diff --git a/.appveyor.yml b/.appveyor.yml
index 5415b8f4a..d4cab45cd 100644
--- a/.appveyor.yml
+++ b/.appveyor.yml
@@ -45,28 +45,27 @@ install:
 
   # Upgrade to the latest version of pip to avoid it displaying warnings
   # about it being out of date.
-  - "pip install --disable-pip-version-check --user --upgrade pip"
+  - "pip install --disable-pip-version-check --user -U pip"
 
   # Install the build dependencies of the project. If some dependencies contain
   # compiled extensions and are not provided as pre-built wheel packages,
   # pip will build them from source using the MSVC compiler matching the
   # target Python version and architecture
-  - "pip install --upgrade setuptools"
-  - "%CMD_IN_ENV% pip install cython fabric fabtools"
-  - "%CMD_IN_ENV% pip install -r requirements.txt"
+  - "%CMD_IN_ENV% python build.py prepare"
 
 build_script:
   # Build the compiled extension
   - "%CMD_IN_ENV% python setup.py build_ext --inplace"
   - ps: appveyor\download.ps1 
   - "tar -xzf corpora/en/wordnet.tar.gz"
-  - "%CMD_IN_ENV% python bin/init_model.py en lang_data/ corpora/ spacy/en/data"
-
+  - "%CMD_IN_ENV% python bin/init_model.py en lang_data/ corpora/ data"
+  - "cp package.json data"
+  - "%CMD_IN_ENV% sputnik build data en_default.sputnik"
+  - "%CMD_IN_ENV% sputnik install en_default.sputnik"
 
 test_script:
   # Run the project tests
-  - "pip install pytest"
-  - "%CMD_IN_ENV% py.test spacy/ -x"
+  - "%CMD_IN_ENV% python build.py test"
 
 after_test:
   # If tests are successful, create binary packages for the project.
diff --git a/.travis.yml b/.travis.yml
index c90da84d2..b6c2a430e 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -1,14 +1,21 @@
 language: python
 
+sudo: required
+dist: precise
+group: edge
+
+python:
+   - "2.7"
+   - "3.4"
+
 os:
   - linux
 
-python:
-  - "2.7"
-  - "3.4"
-  - "3.5"
+env:
+  - PIP_DATE=2015-10-01 MODE=pip
+  - PIP_DATE=2015-10-01 MODE=setup-install
+  - PIP_DATE=2015-10-01 MODE=setup-develop
 
-# install dependencies
 install:
   - "pip install --upgrade setuptools"
   - "pip install cython fabric fabtools"
@@ -21,8 +28,11 @@ install:
   - "mv WordNet-3.0 wordnet"
   - "cd ../../"
   - "export PYTHONPATH=`pwd`"
-  - "python bin/init_model.py en lang_data/ corpora/ spacy/en/data"
+  - "python bin/init_model.py en lang_data/ corpora/ data"
+  - "cp package.json data"
+  - "sputnik build data en_default.sputnik"
+  - "sputnik install en_default.sputnik"
 
-# run tests
 script:
-  - "py.test spacy/ -x"
+  - python build.py $MODE;
+  - python build.py test
diff --git a/MANIFEST.in b/MANIFEST.in
index e69de29bb..18e7a8b29 100644
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -0,0 +1 @@
+recursive-include include *.h
diff --git a/README-MSVC.txt b/README-MSVC.txt
index 2d4a727c1..cc45eab3f 100644
--- a/README-MSVC.txt
+++ b/README-MSVC.txt
@@ -1,14 +1,14 @@
 Python 2.7 Windows build has been tested with the following toolchain:
 	- Python 2.7.10  :)
 	- Microsoft Visual C++ Compiler Package for Python 2.7		http://www.microsoft.com/en-us/download/details.aspx?id=44266
-	- C99 compliant stdint.h for MSVC				http://msinttypes.googlecode.com/svn/trunk/stdint.h
+	- C99 compliant stdint.h for MSVC				https://msinttypes.googlecode.com/svn/trunk/stdint.h
 	 (C99 complian stdint.h header which is not supplied with Microsoft Visual C++ compiler prior to MSVC 2010)
 
 Build steps:
 	- pip install --upgrade setuptools
 	- pip install cython fabric fabtools
 	- pip install -r requirements.txt
-	- python setup.py build_ext --inplace
+	- python pip install -e .
 
 If you are using traditional Microsoft SDK (v7.0 for Python 2.x or v7.1 for Python 3.x) consider run_with_env.cmd from appveyor folder (submodule) as a guideline for environment setup. 
 It can be also used as shell conviguration script for your build, install and run commands, i.e.: cmd /E:ON /V:ON /C run_with_env.cmd <your command> 
diff --git a/README.md b/README.md
index 4fd29e8ef..4f0fc6ad2 100644
--- a/README.md
+++ b/README.md
@@ -45,9 +45,6 @@ Supports
 * OSX
 * Linux 
 * Cygwin
-
-Want to support:
-
 * Visual Studio
 
 Difficult to support:
diff --git a/bin/cythonize.py b/bin/cythonize.py
new file mode 100755
index 000000000..2c18ae8db
--- /dev/null
+++ b/bin/cythonize.py
@@ -0,0 +1,199 @@
+#!/usr/bin/env python
+""" cythonize
+
+Cythonize pyx files into C files as needed.
+
+Usage: cythonize [root_dir]
+
+Default [root_dir] is 'spacy'.
+
+Checks pyx files to see if they have been changed relative to their
+corresponding C files.  If they have, then runs cython on these files to
+recreate the C files.
+
+The script thinks that the pyx files have changed relative to the C files
+by comparing hashes stored in a database file.
+
+Simple script to invoke Cython (and Tempita) on all .pyx (.pyx.in)
+files; while waiting for a proper build system. Uses file hashes to
+figure out if rebuild is needed.
+
+For now, this script should be run by developers when changing Cython files
+only, and the resulting C files checked in, so that end-users (and Python-only
+developers) do not get the Cython/Tempita dependencies.
+
+Originally written by Dag Sverre Seljebotn, and copied here from:
+
+https://raw.github.com/dagss/private-scipy-refactor/cythonize/cythonize.py
+
+Note: this script does not check any of the dependent C libraries; it only
+operates on the Cython .pyx files.
+"""
+
+from __future__ import division, print_function, absolute_import
+
+import os
+import re
+import sys
+import hashlib
+import subprocess
+
+HASH_FILE = 'cythonize.dat'
+DEFAULT_ROOT = 'spacy'
+VENDOR = 'spaCy'
+
+# WindowsError is not defined on unix systems
+try:
+    WindowsError
+except NameError:
+    WindowsError = None
+
+#
+# Rules
+#
+def process_pyx(fromfile, tofile):
+    try:
+        from Cython.Compiler.Version import version as cython_version
+        from distutils.version import LooseVersion
+        if LooseVersion(cython_version) < LooseVersion('0.19'):
+            raise Exception('Building %s requires Cython >= 0.19' % VENDOR)
+
+    except ImportError:
+        pass
+
+    flags = ['--fast-fail']
+    if tofile.endswith('.cpp'):
+        flags += ['--cplus']
+
+    try:
+        try:
+            r = subprocess.call(['cython'] + flags + ["-o", tofile, fromfile])
+            if r != 0:
+                raise Exception('Cython failed')
+        except OSError:
+            # There are ways of installing Cython that don't result in a cython
+            # executable on the path, see gh-2397.
+            r = subprocess.call([sys.executable, '-c',
+                                 'import sys; from Cython.Compiler.Main import '
+                                 'setuptools_main as main; sys.exit(main())'] + flags +
+                                 ["-o", tofile, fromfile])
+            if r != 0:
+                raise Exception('Cython failed')
+    except OSError:
+        raise OSError('Cython needs to be installed')
+
+def process_tempita_pyx(fromfile, tofile):
+    try:
+        try:
+            from Cython import Tempita as tempita
+        except ImportError:
+            import tempita
+    except ImportError:
+        raise Exception('Building %s requires Tempita: '
+                        'pip install --user Tempita' % VENDOR)
+    with open(fromfile, "r") as f:
+        tmpl = f.read()
+    pyxcontent = tempita.sub(tmpl)
+    assert fromfile.endswith('.pyx.in')
+    pyxfile = fromfile[:-len('.pyx.in')] + '.pyx'
+    with open(pyxfile, "w") as f:
+        f.write(pyxcontent)
+    process_pyx(pyxfile, tofile)
+
+rules = {
+    # fromext : function
+    '.pyx' : process_pyx,
+    '.pyx.in' : process_tempita_pyx
+    }
+#
+# Hash db
+#
+def load_hashes(filename):
+    # Return { filename : (sha1 of input, sha1 of output) }
+    if os.path.isfile(filename):
+        hashes = {}
+        with open(filename, 'r') as f:
+            for line in f:
+                filename, inhash, outhash = line.split()
+                hashes[filename] = (inhash, outhash)
+    else:
+        hashes = {}
+    return hashes
+
+def save_hashes(hash_db, filename):
+    with open(filename, 'w') as f:
+        for key, value in sorted(hash_db.items()):
+            f.write("%s %s %s\n" % (key, value[0], value[1]))
+
+def sha1_of_file(filename):
+    h = hashlib.sha1()
+    with open(filename, "rb") as f:
+        h.update(f.read())
+    return h.hexdigest()
+
+#
+# Main program
+#
+
+def normpath(path):
+    path = path.replace(os.sep, '/')
+    if path.startswith('./'):
+        path = path[2:]
+    return path
+
+def get_hash(frompath, topath):
+    from_hash = sha1_of_file(frompath)
+    to_hash = sha1_of_file(topath) if os.path.exists(topath) else None
+    return (from_hash, to_hash)
+
+def process(path, fromfile, tofile, processor_function, hash_db):
+    fullfrompath = os.path.join(path, fromfile)
+    fulltopath = os.path.join(path, tofile)
+    current_hash = get_hash(fullfrompath, fulltopath)
+    if current_hash == hash_db.get(normpath(fullfrompath), None):
+        print('%s has not changed' % fullfrompath)
+        return
+
+    orig_cwd = os.getcwd()
+    try:
+        os.chdir(path)
+        print('Processing %s' % fullfrompath)
+        processor_function(fromfile, tofile)
+    finally:
+        os.chdir(orig_cwd)
+    # changed target file, recompute hash
+    current_hash = get_hash(fullfrompath, fulltopath)
+    # store hash in db
+    hash_db[normpath(fullfrompath)] = current_hash
+
+
+def find_process_files(root_dir):
+    hash_db = load_hashes(HASH_FILE)
+    for cur_dir, dirs, files in os.walk(root_dir):
+        for filename in files:
+            in_file = os.path.join(cur_dir, filename + ".in")
+            if filename.endswith('.pyx') and os.path.isfile(in_file):
+                continue
+            for fromext, function in rules.items():
+                if filename.endswith(fromext):
+                    toext = ".cpp"
+                    # with open(os.path.join(cur_dir, filename), 'rb') as f:
+                    #     data = f.read()
+                    #     m = re.search(br"^\s*#\s*distutils:\s*language\s*=\s*c\+\+\s*$", data, re.I|re.M)
+                    #     if m:
+                    #         toext = ".cxx"
+                    fromfile = filename
+                    tofile = filename[:-len(fromext)] + toext
+                    process(cur_dir, fromfile, tofile, function, hash_db)
+                    save_hashes(hash_db, HASH_FILE)
+
+def main():
+    try:
+        root_dir = sys.argv[1]
+    except IndexError:
+        root_dir = DEFAULT_ROOT
+    find_process_files(root_dir)
+
+
+if __name__ == '__main__':
+    main()
\ No newline at end of file
diff --git a/build.py b/build.py
new file mode 100644
index 000000000..ac3c489e3
--- /dev/null
+++ b/build.py
@@ -0,0 +1,71 @@
+#!/usr/bin/env python
+from __future__ import print_function
+import os
+import sys
+import shutil
+from subprocess import call
+
+
+def x(cmd):
+    print('$ '+cmd)
+    res = call(cmd, shell=True)
+    if res != 0:
+        sys.exit(res)
+
+
+if len(sys.argv) < 2:
+    print('usage: %s <install-mode> [<pip-date>]')
+    sys.exit(1)
+
+
+install_mode = sys.argv[1]
+
+
+if install_mode == 'prepare':
+    x('python pip-clear.py')
+
+    pip_date = len(sys.argv) > 2 and sys.argv[2]
+    if pip_date:
+        x('python pip-date.py %s pip setuptools wheel six' % pip_date)
+
+    x('pip install -r requirements.txt')
+    x('pip list')
+
+
+elif install_mode == 'pip':
+    if os.path.exists('dist'):
+        shutil.rmtree('dist')
+    x('python setup.py sdist')
+    x('python pip-clear.py')
+
+    filenames = os.listdir('dist')
+    assert len(filenames) == 1
+    x('pip list')
+    x('pip install dist/%s' % filenames[0])
+
+
+elif install_mode == 'setup-install':
+    x('python setup.py install')
+
+
+elif install_mode == 'setup-develop':
+    x('pip install -e .')
+
+
+elif install_mode == 'test':
+    x('pip install pytest')
+    x('pip list')
+
+    if os.path.exists('tmp'):
+        shutil.rmtree('tmp')
+    os.mkdir('tmp')
+
+    try:
+        old = os.getcwd()
+        os.chdir('tmp')
+
+        x('python -m spacy.en.download')
+        x('python -m pytest --tb="native" ../spacy/ -x --models --vectors --slow')
+
+    finally:
+        os.chdir(old)
diff --git a/examples/pos_tag.py b/examples/pos_tag.py
new file mode 100644
index 000000000..a118966b0
--- /dev/null
+++ b/examples/pos_tag.py
@@ -0,0 +1,89 @@
+'''Print part-of-speech tagged, true-cased, (very roughly) sentence-separated
+text, with each "sentence" on a newline, and spaces between tokens. Supports
+multi-processing.
+'''
+from __future__ import print_function, unicode_literals, division
+import io
+import bz2
+import logging
+from toolz import partition
+from os import path
+
+import spacy.en
+
+from joblib import Parallel, delayed
+import plac
+import ujson
+
+
+def parallelize(func, iterator, n_jobs, extra):
+    extra = tuple(extra)
+    return Parallel(n_jobs=n_jobs)(delayed(func)(*(item + extra)) for item in iterator)
+
+
+def iter_texts_from_json_bz2(loc):
+    '''
+    Iterator of unicode strings, one per document (here, a comment).
+    
+    Expects a a path to a BZ2 file, which should be new-line delimited JSON. The
+    document text should be in a string field titled 'body'.
+
+    This is the data format of the Reddit comments corpus.
+    '''
+    with bz2.BZ2File(loc) as file_:
+        for i, line in enumerate(file_):
+            yield ujson.loads(line)['body']
+
+
+def transform_texts(batch_id, input_, out_dir):
+    out_loc = path.join(out_dir, '%d.txt' % batch_id)
+    if path.exists(out_loc):
+        return None
+    print('Batch', batch_id)
+    nlp = spacy.en.English(parser=False, entity=False)
+    with io.open(out_loc, 'w', encoding='utf8') as file_:
+        for text in input_:
+            doc = nlp(text)
+            file_.write(' '.join(represent_word(w) for w in doc if not w.is_space))
+            file_.write('\n')
+
+
+def represent_word(word):
+    text = word.text
+    # True-case, i.e. try to normalize sentence-initial capitals.
+    # Only do this if the lower-cased form is more probable.
+    if text.istitle() \
+    and is_sent_begin(word) \
+    and word.prob < word.vocab[text.lower()].prob:
+        text = text.lower()
+    return text + '|' + word.tag_
+
+
+def is_sent_begin(word):
+    # It'd be nice to have some heuristics like these in the library, for these
+    # times where we don't care so much about accuracy of SBD, and we don't want
+    # to parse
+    if word.i == 0:
+        return True
+    elif word.i >= 2 and word.nbor(-1).text in ('.', '!', '?', '...'):
+        return True
+    else:
+        return False
+
+
+@plac.annotations(
+    in_loc=("Location of input file"),
+    out_dir=("Location of input file"),
+    n_workers=("Number of workers", "option", "n", int),
+    batch_size=("Batch-size for each process", "option", "b", int)
+)
+def main(in_loc, out_dir, n_workers=4, batch_size=100000):
+    if not path.exists(out_dir):
+        path.join(out_dir)
+    texts = partition(batch_size, iter_texts(in_loc))
+    parallelize(transform_texts, enumerate(texts), n_workers, [out_dir])
+ 
+
+if __name__ == '__main__':
+    plac.call(main)
+
diff --git a/fabfile.py b/fabfile.py
index 85c355b3e..7a53fb88f 100644
--- a/fabfile.py
+++ b/fabfile.py
@@ -74,7 +74,6 @@ def web():
     jade('home/index.jade', '')
     jade('docs/index.jade', 'docs/')
     jade('blog/index.jade', 'blog/')
-    jade('tutorials/index.jade', 'tutorials/')
 
     for collection in ('blog', 'tutorials'):
         for post_dir in (Path(__file__).parent / 'website' / 'src' / 'jade' / collection).iterdir():
@@ -85,7 +84,50 @@ def web():
 
 
 def web_publish(assets_path):
-    local('aws s3 sync --delete --exclude "resources/*" website/site/ s3://spacy.io')
+    from boto.s3.connection import S3Connection, OrdinaryCallingFormat
+
+    site_path = 'website/site'
+
+    os.environ['S3_USE_SIGV4'] = 'True'
+    conn = S3Connection(host='s3.eu-central-1.amazonaws.com',
+                        calling_format=OrdinaryCallingFormat())
+    bucket = conn.get_bucket('spacy.io', validate=False)
+
+    keys_left = set([k.name for k in bucket.list()
+                     if not k.name.startswith('resources')])
+
+    for root, dirnames, filenames in os.walk(site_path):
+        for dirname in dirnames:
+            target = os.path.relpath(os.path.join(root, dirname), site_path)
+            source = os.path.join(target, 'index.html')
+
+            if os.path.exists(os.path.join(root, dirname, 'index.html')):
+                key = bucket.new_key(source)
+                key.set_redirect('//%s/%s' % (bucket.name, target))
+                print('adding redirect for %s' % target)
+
+                keys_left.remove(source)
+
+        for filename in filenames:
+            source = os.path.join(root, filename)
+
+            target = os.path.relpath(root, site_path)
+            if target == '.':
+                target = filename
+            elif filename != 'index.html':
+                target = os.path.join(target, filename)
+
+            key = bucket.new_key(target)
+            key.set_metadata('Content-Type', 'text/html')
+            key.set_contents_from_filename(source)
+            print('uploading %s' % target)
+
+            keys_left.remove(target)
+
+    for key_name in keys_left:
+        print('deleting %s' % key_name)
+        bucket.delete_key(key_name)
+
     local('aws s3 sync --delete %s s3://spacy.io/resources' % assets_path)
 
 
diff --git a/package.json b/package.json
new file mode 100644
index 000000000..0710031ed
--- /dev/null
+++ b/package.json
@@ -0,0 +1,17 @@
+{
+    "name": "en_default",
+    "version": "0.100.0",
+    "description": "english default model",
+    "license": "public domain",
+    "include": [
+        "deps/*",
+        "ner/*",
+        "pos/*",
+        "tokenizer/*",
+        "vocab/*",
+        "wordnet/*"
+    ],
+    "compatibility": {
+        "spacy": "==0.100.0"
+    }
+}
diff --git a/pip-clear.py b/pip-clear.py
new file mode 100755
index 000000000..573a20dcd
--- /dev/null
+++ b/pip-clear.py
@@ -0,0 +1,28 @@
+#!/usr/bin/env python
+from __future__ import print_function
+
+from pip.commands.uninstall import UninstallCommand
+from pip import get_installed_distributions
+
+
+packages = []
+for package in get_installed_distributions():
+    if package.location.endswith('dist-packages'):
+        continue
+    elif package.project_name in ('pip', 'setuptools'):
+        continue
+    packages.append(package.project_name)
+
+
+if packages:
+    pip = UninstallCommand()
+    options, args = pip.parse_args(packages)
+    options.yes = True
+
+    try:
+        pip.run(options, args)
+    except OSError as e:
+        if e.errno != 13:
+            raise e
+        print("You lack permissions to uninstall this package. Perhaps run with sudo? Exiting.")
+        exit(13)
diff --git a/pip-date.py b/pip-date.py
new file mode 100644
index 000000000..2a775255a
--- /dev/null
+++ b/pip-date.py
@@ -0,0 +1,96 @@
+#!/usr/bin/env python
+from __future__ import print_function
+import json
+import re
+import sys
+from bisect import bisect
+from datetime import datetime
+from datetime import timedelta
+import ssl
+
+try:
+    from urllib.request import Request, build_opener, HTTPSHandler, URLError
+except ImportError:
+    from urllib2 import Request, build_opener, HTTPSHandler, URLError
+
+from pip.commands.uninstall import UninstallCommand
+from pip.commands.install import InstallCommand
+from pip import get_installed_distributions
+
+
+def get_releases(package_name):
+    url = 'https://pypi.python.org/pypi/%s/json' % package_name
+
+    ssl_context = HTTPSHandler(
+        context=ssl.SSLContext(ssl.PROTOCOL_TLSv1))
+    opener = build_opener(ssl_context)
+
+    retries = 10
+    while retries > 0:
+        try:
+            r = opener.open(Request(url))
+            break
+        except URLError:
+            retries -= 1
+
+    return json.loads(r.read().decode('utf8'))['releases']
+
+
+def parse_iso8601(s):
+    return datetime(*map(int, re.split('[^\d]', s)))
+
+
+def select_version(select_date, package_name):
+    versions = []
+    for version, dists in get_releases(package_name).items():
+        date = [parse_iso8601(d['upload_time']) for d in dists]
+        if date:
+            versions.append((sorted(date)[0], version))
+
+    versions = sorted(versions)
+    min_date = versions[0][0]
+    if select_date < min_date:
+        raise Exception('invalid select_date: %s, must be '
+                        '%s or newer.' % (select_date, min_date))
+
+    return versions[bisect([x[0] for x in versions], select_date) - 1][1]
+
+
+installed_packages = [
+    package.project_name
+    for package in
+    get_installed_distributions()
+    if (not package.location.endswith('dist-packages') and
+        package.project_name not in ('pip', 'setuptools'))
+]
+
+if installed_packages:
+    pip = UninstallCommand()
+    options, args = pip.parse_args(installed_packages)
+    options.yes = True
+
+    try:
+        pip.run(options, args)
+    except OSError as e:
+        if e.errno != 13:
+            raise e
+        print("You lack permissions to uninstall this package. Perhaps run with sudo? Exiting.")
+        exit(13)
+
+
+date = parse_iso8601(sys.argv[1])
+packages = {p: select_version(date, p) for p in sys.argv[2:]}
+args = ['=='.join(a) for a in packages.items()]
+
+cmd = InstallCommand()
+options, args = cmd.parse_args(args)
+options.ignore_installed = True
+options.force_reinstall = True
+
+try:
+    print(cmd.run(options, args))
+except OSError as e:
+    if e.errno != 13:
+        raise e
+    print("You lack permissions to uninstall this package. Perhaps run with sudo? Exiting.")
+    exit(13)
diff --git a/requirements.txt b/requirements.txt
index 344cc7665..9a125407d 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,13 +1,13 @@
 cython
 cymem == 1.30
 pathlib
-preshed == 0.44
-thinc == 4.0
-murmurhash == 0.24
+preshed == 0.46.1
+thinc == 4.1.0
+murmurhash == 0.26
 text-unidecode
 numpy
 plac
 six
 ujson
 cloudpickle
-sputnik == 0.5.2
+sputnik == 0.6.4
diff --git a/setup.py b/setup.py
index 8e3e72087..53290138a 100644
--- a/setup.py
+++ b/setup.py
@@ -1,231 +1,281 @@
 #!/usr/bin/env python
-from setuptools import setup
-import shutil
-
-import sys
+from __future__ import division, print_function
 import os
-from os import path
-
-from setuptools import Extension
-from distutils import sysconfig
-from distutils.core import setup, Extension
+import shutil
+import subprocess
+import sys
+import contextlib
 from distutils.command.build_ext import build_ext
+from distutils.sysconfig import get_python_inc
 
-import platform
+try:
+    from setuptools import Extension, setup
+except ImportError:
+    from distutils.core import Extension, setup
 
-PACKAGE_DATA =  {
-    "spacy": ["*.pxd"],
-    "spacy.tokens": ["*.pxd"],
-    "spacy.serialize": ["*.pxd"],
-    "spacy.syntax": ["*.pxd"],
-    "spacy.en": [
-        "*.pxd",
-        "data/wordnet/*.exc",
-        "data/wordnet/index.*",
-        "data/tokenizer/*",
-        "data/vocab/serializer.json"
-    ]
-}
+
+MAJOR      = 0
+MINOR      = 100
+MICRO      = 0
+ISRELEASED = False
+VERSION    = '%d.%d.%d' % (MAJOR, MINOR, MICRO)
+
+
+PACKAGES = [
+    'spacy',
+    'spacy.tokens',
+    'spacy.en',
+    'spacy.serialize',
+    'spacy.syntax',
+    'spacy.munge',
+    'spacy.tests',
+    'spacy.tests.matcher',
+    'spacy.tests.morphology',
+    'spacy.tests.munge',
+    'spacy.tests.parser',
+    'spacy.tests.serialize',
+    'spacy.tests.spans',
+    'spacy.tests.tagger',
+    'spacy.tests.tokenizer',
+    'spacy.tests.tokens',
+    'spacy.tests.vectors',
+    'spacy.tests.vocab']
+
+
+MOD_NAMES = [
+    'spacy.parts_of_speech',
+    'spacy.strings',
+    'spacy.lexeme',
+    'spacy.vocab',
+    'spacy.attrs',
+    'spacy.morphology',
+    'spacy.tagger',
+    'spacy.syntax.stateclass',
+    'spacy.tokenizer',
+    'spacy.syntax.parser',
+    'spacy.syntax.transition_system',
+    'spacy.syntax.arc_eager',
+    'spacy.syntax._parse_features',
+    'spacy.gold',
+    'spacy.orth',
+    'spacy.tokens.doc',
+    'spacy.tokens.span',
+    'spacy.tokens.token',
+    'spacy.serialize.packer',
+    'spacy.serialize.huffman',
+    'spacy.serialize.bits',
+    'spacy.cfile',
+    'spacy.matcher',
+    'spacy.syntax.ner',
+    'spacy.symbols']
+
+
+if sys.version_info[:2] < (2, 7) or (3, 0) <= sys.version_info[0:2] < (3, 4):
+    raise RuntimeError('Python version 2.7 or >= 3.4 required.')
 
 
 # By subclassing build_extensions we have the actual compiler that will be used which is really known only after finalize_options
 # http://stackoverflow.com/questions/724664/python-distutils-how-to-get-a-compiler-that-is-going-to-be-used
-compile_options =  {'msvc'  : ['/Ox', '/EHsc']  ,
-                    'other' : ['-O3', '-Wno-strict-prototypes', '-Wno-unused-function']       }
-link_options    =  {'msvc'  : [] ,
-                    'other' : [] }
+compile_options =  {'msvc'  : ['/Ox', '/EHsc'],
+                    'other' : ['-O3', '-Wno-strict-prototypes', '-Wno-unused-function']}
+link_options    =  {'msvc'  : [],
+                    'other' : []}
+
+if sys.platform.startswith('darwin'):
+    compile_options['other'].append('-mmacosx-version-min=10.8')
+    compile_options['other'].append('-stdlib=libc++')
+    link_options['other'].append('-lc++')
+
+
 class build_ext_options:
     def build_options(self):
-        c_type = None
-        if self.compiler.compiler_type in compile_options:
-            c_type = self.compiler.compiler_type
-        elif 'other' in compile_options:
-            c_type = 'other'
-        if c_type is not None:
-           for e in self.extensions:
-               e.extra_compile_args = compile_options[c_type]
+        for e in self.extensions:
+            e.extra_compile_args = compile_options.get(
+                self.compiler.compiler_type, compile_options['other'])
+        for e in self.extensions:
+            e.extra_link_args = link_options.get(
+                self.compiler.compiler_type, link_options['other'])
 
-        l_type = None 
-        if self.compiler.compiler_type in link_options:
-            l_type = self.compiler.compiler_type
-        elif 'other' in link_options:
-            l_type = 'other'
-        if l_type is not None:
-           for e in self.extensions:
-               e.extra_link_args = link_options[l_type]
 
-class build_ext_subclass( build_ext, build_ext_options ):
+class build_ext_subclass(build_ext, build_ext_options):
     def build_extensions(self):
         build_ext_options.build_options(self)
         build_ext.build_extensions(self)
-        
-    
-
-# PyPy --- NB! PyPy doesn't really work, it segfaults all over the place. But,
-# this is necessary to get it compile.
-# We have to resort to monkey-patching to set the compiler, because pypy broke
-# all the everything.
-
-pre_patch_customize_compiler = sysconfig.customize_compiler
-def my_customize_compiler(compiler):
-    pre_patch_customize_compiler(compiler)
-    compiler.compiler_cxx = ['c++']
 
 
-if platform.python_implementation() == 'PyPy':
-    sysconfig.customize_compiler = my_customize_compiler
+# Return the git revision as a string
+def git_version():
+    def _minimal_ext_cmd(cmd):
+        # construct minimal environment
+        env = {}
+        for k in ['SYSTEMROOT', 'PATH']:
+            v = os.environ.get(k)
+            if v is not None:
+                env[k] = v
+        # LANGUAGE is used on win32
+        env['LANGUAGE'] = 'C'
+        env['LANG'] = 'C'
+        env['LC_ALL'] = 'C'
+        out = subprocess.Popen(cmd, stdout = subprocess.PIPE, env=env).communicate()[0]
+        return out
 
-#def install_headers():
-#    dest_dir = path.join(sys.prefix, 'include', 'murmurhash')
-#    if not path.exists(dest_dir):
-#        shutil.copytree('murmurhash/headers/murmurhash', dest_dir)
-#
-#    dest_dir = path.join(sys.prefix, 'include', 'numpy')
+    try:
+        out = _minimal_ext_cmd(['git', 'rev-parse', 'HEAD'])
+        GIT_REVISION = out.strip().decode('ascii')
+    except OSError:
+        GIT_REVISION = 'Unknown'
+
+    return GIT_REVISION
 
 
-includes = ['.', path.join(sys.prefix, 'include')]
+def get_version_info():
+    # Adding the git rev number needs to be done inside write_version_py(),
+    # otherwise the import of spacy.about messes up the build under Python 3.
+    FULLVERSION = VERSION
+    if os.path.exists('.git'):
+        GIT_REVISION = git_version()
+    elif os.path.exists(os.path.join('spacy', 'about.py')):
+        # must be a source distribution, use existing version file
+        try:
+            from spacy.about import git_revision as GIT_REVISION
+        except ImportError:
+            raise ImportError('Unable to import git_revision. Try removing '
+                              'spacy/about.py and the build directory '
+                              'before building.')
+    else:
+        GIT_REVISION = 'Unknown'
+
+    if not ISRELEASED:
+        FULLVERSION += '.dev0+' + GIT_REVISION[:7]
+
+    return FULLVERSION, GIT_REVISION
 
 
-try:
-    import numpy
-    numpy_headers = path.join(numpy.get_include(), 'numpy')
-    shutil.copytree(numpy_headers, path.join(sys.prefix, 'include', 'numpy'))
-except ImportError:
-    pass
-except OSError:
-    pass
+def write_version(path):
+    cnt = """# THIS FILE IS GENERATED FROM SPACY SETUP.PY
+short_version = '%(version)s'
+version = '%(version)s'
+full_version = '%(full_version)s'
+git_revision = '%(git_revision)s'
+release = %(isrelease)s
+if not release:
+    version = full_version
+"""
+    FULLVERSION, GIT_REVISION = get_version_info()
+
+    with open(path, 'w') as f:
+        f.write(cnt % {'version': VERSION,
+                       'full_version' : FULLVERSION,
+                       'git_revision' : GIT_REVISION,
+                       'isrelease': str(ISRELEASED)})
 
 
+def generate_cython(root, source):
+    print('Cythonizing sources')
+    p = subprocess.call([sys.executable,
+                         os.path.join(root, 'bin', 'cythonize.py'),
+                         source])
+    if p != 0:
+        raise RuntimeError('Running cythonize failed')
 
 
-def clean(mod_names):
-    for name in mod_names:
+def import_include(module_name):
+    try:
+        return __import__(module_name, globals(), locals(), [], 0)
+    except ImportError:
+        raise ImportError('Unable to import %s. Create a virtual environment '
+                          'and install all dependencies from requirements.txt, '
+                          'e.g., run "pip install -r requirements.txt".' % module_name)
+
+
+def copy_include(src, dst, path):
+    assert os.path.isdir(src)
+    assert os.path.isdir(dst)
+    shutil.copytree(
+        os.path.join(src, path),
+        os.path.join(dst, path))
+
+
+def prepare_includes(path):
+    include_dir = os.path.join(path, 'include')
+    if os.path.exists(include_dir):
+        shutil.rmtree(include_dir)
+    os.mkdir(include_dir)
+
+    numpy = import_include('numpy')
+    copy_include(numpy.get_include(), include_dir, 'numpy')
+
+    murmurhash = import_include('murmurhash')
+    copy_include(murmurhash.get_include(), include_dir, 'murmurhash')
+
+
+def is_source_release(path):
+    return os.path.exists(os.path.join(path, 'PKG-INFO'))
+
+
+def clean(path):
+    for name in MOD_NAMES:
         name = name.replace('.', '/')
-        so = name + '.so'
-        html = name + '.html'
-        cpp = name + '.cpp'
-        c = name + '.c'
-        for file_path in [so, html, cpp, c]:
+        for ext in ['.so', '.html', '.cpp', '.c']:
+            file_path = os.path.join(path, name + ext)
             if os.path.exists(file_path):
                 os.unlink(file_path)
 
 
-def name_to_path(mod_name, ext):
-    return '%s.%s' % (mod_name.replace('.', '/'), ext)
+@contextlib.contextmanager
+def chdir(new_dir):
+    old_dir = os.getcwd()
+    try:
+        os.chdir(new_dir)
+        sys.path.insert(0, new_dir)
+        yield
+    finally:
+        del sys.path[0]
+        os.chdir(old_dir)
 
 
-def c_ext(mod_name, language, includes):
-    mod_path = name_to_path(mod_name, language)
-    return Extension(mod_name, [mod_path], include_dirs=includes)
+def setup_package():
+    root = os.path.abspath(os.path.dirname(__file__))
 
+    if len(sys.argv) > 1 and sys.argv[1] == 'clean':
+        return clean(root)
 
-def cython_setup(mod_names, language, includes):
-    import Cython.Distutils
-    import Cython.Build
-    import distutils.core
+    with chdir(root):
+        write_version(os.path.join(root, 'spacy', 'about.py'))
 
-    class build_ext_cython_subclass( Cython.Distutils.build_ext, build_ext_options ):
-        def build_extensions(self):
-            build_ext_options.build_options(self)
-            Cython.Distutils.build_ext.build_extensions(self)
+        include_dirs = [
+            get_python_inc(plat_specific=True),
+            os.path.join(root, 'include')]
 
-    if language == 'cpp':
-        language = 'c++'
-    exts = []
-    for mod_name in mod_names:
-        mod_path = mod_name.replace('.', '/') + '.pyx'
-        e = Extension(mod_name, [mod_path], language=language, include_dirs=includes)
-        exts.append(e)
-    distutils.core.setup(
-        name='spacy',
-        packages=['spacy', 'spacy.tokens', 'spacy.en', 'spacy.serialize',
-                  'spacy.syntax', 'spacy.munge'],
-        description="Industrial-strength NLP",
-        author='Matthew Honnibal',
-        author_email='honnibal@gmail.com',
-        version=VERSION,
-        url="http://spacy.io",
-        package_data=PACKAGE_DATA,
-        ext_modules=exts,
-        cmdclass={'build_ext': build_ext_cython_subclass},
-        license="MIT",
-    )
+        ext_modules = []
+        for mod_name in MOD_NAMES:
+            mod_path = mod_name.replace('.', '/') + '.cpp'
+            ext_modules.append(
+                Extension(mod_name, [mod_path],
+                    language='c++', include_dirs=include_dirs))
 
+        if not is_source_release(root):
+            generate_cython(root, 'spacy')
+            prepare_includes(root)
 
-def run_setup(exts):
-    setup(
-        name='spacy',
-        packages=['spacy', 'spacy.tokens', 'spacy.en', 'spacy.serialize',
-                  'spacy.syntax', 'spacy.munge',
-                  'spacy.tests',
-                  'spacy.tests.matcher',
-                  'spacy.tests.morphology',
-                  'spacy.tests.munge',
-                  'spacy.tests.parser',
-                  'spacy.tests.serialize',
-                  'spacy.tests.spans',
-                  'spacy.tests.tagger',
-                  'spacy.tests.tokenizer',
-                  'spacy.tests.tokens',
-                  'spacy.tests.vectors',
-                  'spacy.tests.vocab'],
-        description="Industrial-strength NLP",
-        author='Matthew Honnibal',
-        author_email='honnibal@gmail.com',
-        version=VERSION,
-        url="http://honnibal.github.io/spaCy/",
-        package_data=PACKAGE_DATA,
-        ext_modules=exts,
-        license="MIT",
-        install_requires=['numpy', 'murmurhash == 0.24', 'cymem == 1.30', 'preshed == 0.44',
-                          'thinc == 4.0.0', "text_unidecode", 'plac', 'six',
-                          'ujson', 'cloudpickle', 'sputnik == 0.5.2'],
-        setup_requires=["headers_workaround"],
-        cmdclass = {'build_ext': build_ext_subclass },
-    )
-
-    import headers_workaround
-
-    headers_workaround.fix_venv_pypy_include()
-    headers_workaround.install_headers('murmurhash')
-    headers_workaround.install_headers('numpy')
-
-
-VERSION = '0.100'
-def main(modules, is_pypy):
-    language = "cpp"
-    includes = ['.', path.join(sys.prefix, 'include')]
-    if sys.platform.startswith('darwin'):
-        compile_options['other'].append('-mmacosx-version-min=10.8')
-        compile_options['other'].append('-stdlib=libc++')
-        link_options['other'].append('-lc++')
-    if use_cython:
-        cython_setup(modules, language, includes)
-    else:
-        exts = [c_ext(mn, language, includes)
-                      for mn in modules]
-        run_setup(exts)
-
-MOD_NAMES = ['spacy.parts_of_speech', 'spacy.strings',
-             'spacy.lexeme', 'spacy.vocab', 'spacy.attrs',
-             'spacy.morphology', 'spacy.tagger',
-             'spacy.syntax.stateclass', 
-             'spacy.tokenizer',
-             'spacy.syntax.parser', 
-             'spacy.syntax.transition_system',
-             'spacy.syntax.arc_eager',
-             'spacy.syntax._parse_features',
-             'spacy.gold', 'spacy.orth',
-             'spacy.tokens.doc', 'spacy.tokens.span', 'spacy.tokens.token',
-             'spacy.serialize.packer', 'spacy.serialize.huffman', 'spacy.serialize.bits',
-             'spacy.cfile', 'spacy.matcher',
-             'spacy.syntax.ner',
-             'spacy.symbols']
+        setup(
+            name='spacy',
+            packages=PACKAGES,
+            package_data={'': ['*.pyx', '*.pxd']},
+            description='Industrial-strength NLP',
+            author='Matthew Honnibal',
+            author_email='matt@spacy.io',
+            version=VERSION,
+            url='https://spacy.io',
+            license='MIT',
+            ext_modules=ext_modules,
+            install_requires=['numpy', 'murmurhash == 0.26', 'cymem == 1.30', 'preshed == 0.46.1',
+                              'thinc == 4.1.0', 'text_unidecode', 'plac', 'six',
+                              'ujson', 'cloudpickle', 'sputnik == 0.6.4'],
+            cmdclass = {
+                'build_ext': build_ext_subclass},
+        )
 
 
 if __name__ == '__main__':
-    if sys.argv[1] == 'clean':
-        clean(MOD_NAMES)
-    else:
-        use_cython = sys.argv[1] == 'build_ext'
-        main(MOD_NAMES, use_cython)
+    setup_package()
diff --git a/spacy/_nn.py b/spacy/_nn.py
deleted file mode 100644
index 48dca390c..000000000
--- a/spacy/_nn.py
+++ /dev/null
@@ -1,3 +0,0 @@
-"""Feed-forward neural network, using Thenao."""
-
-
diff --git a/spacy/_nn.pyx b/spacy/_nn.pyx
deleted file mode 100644
index c47be1f49..000000000
--- a/spacy/_nn.pyx
+++ /dev/null
@@ -1,146 +0,0 @@
-"""Feed-forward neural network, using Thenao."""
-
-import os
-import sys
-import time
-
-import numpy
-
-import theano
-import theano.tensor as T
-import plac
-
-from spacy.gold import read_json_file
-from spacy.gold import GoldParse
-from spacy.en.pos import POS_TEMPLATES, POS_TAGS, setup_model_dir
-
-
-def build_model(n_classes, n_vocab, n_hidden, n_word_embed, n_tag_embed):
-    # allocate symbolic variables for the data
-    words = T.vector('words')
-    tags = T.vector('tags') 
-    
-    word_e = _init_embedding(n_words, n_word_embed)
-    tag_e = _init_embedding(n_tags, n_tag_embed)
-    label_e = _init_embedding(n_labels, n_label_embed)
-    maxent_W, maxent_b = _init_maxent_weights(n_hidden, n_classes)
-    hidden_W, hidden_b = _init_hidden_weights(28*28, n_hidden, T.tanh) 
-    params = [hidden_W, hidden_b, maxent_W, maxent_b, word_e, tag_e, label_e]
-
-    x = T.concatenate([
-          T.flatten(word_e[word_indices], outdim=1),
-          T.flatten(tag_e[tag_indices], outdim=1)])
-
-    p_y_given_x = feed_layer(
-                    T.nnet.softmax,
-                    maxent_W,
-                    maxent_b,
-                      feed_layer(
-                        T.tanh,
-                        hidden_W,
-                        hidden_b,
-                        x))[0]
-
-    guess = T.argmax(p_y_given_x)
-
-    cost = (
-        -T.log(p_y_given_x[y])
-        + L1(L1_reg, maxent_W, hidden_W, word_e, tag_e)
-        + L2(L2_reg, maxent_W, hidden_W, wod_e, tag_e)
-    )
-
-    train_model = theano.function(
-        inputs=[words, tags, y],
-        outputs=guess,
-        updates=[update(learning_rate, param, cost) for param in params]
-    )
-
-    evaluate_model = theano.function(
-        inputs=[x, y],
-        outputs=T.neq(y, T.argmax(p_y_given_x[0])),
-    )
-    return train_model, evaluate_model
-
-
-def _init_embedding(vocab_size, n_dim):
-    embedding = 0.2 * numpy.random.uniform(-1.0, 1.0, (vocab_size+1, n_dim))
-    return theano.shared(embedding).astype(theano.config.floatX)
-
-
-def _init_maxent_weights(n_hidden, n_out):
-    weights = numpy.zeros((n_hidden, 10), dtype=theano.config.floatX)
-    bias =  numpy.zeros((10,), dtype=theano.config.floatX)
-    return (
-        theano.shared(name='W', borrow=True, value=weights),
-        theano.shared(name='b', borrow=True, value=bias)
-    )
-
-
-def _init_hidden_weights(n_in, n_out, activation=T.tanh):
-    rng = numpy.random.RandomState(1234)
-    weights = numpy.asarray(
-        rng.uniform(
-            low=-numpy.sqrt(6. / (n_in + n_out)),
-            high=numpy.sqrt(6. / (n_in + n_out)),
-            size=(n_in, n_out)
-        ),
-        dtype=theano.config.floatX
-    )
-
-    bias = numpy.zeros((n_out,), dtype=theano.config.floatX)
-    return (
-        theano.shared(value=weights, name='W', borrow=True),
-        theano.shared(value=bias, name='b', borrow=True)
-    )
-
-           
-def feed_layer(activation, weights, bias, input):
-    return activation(T.dot(input, weights) + bias)
-
-
-def L1(L1_reg, w1, w2):
-    return L1_reg * (abs(w1).sum() + abs(w2).sum())
-
-
-def L2(L2_reg, w1, w2):
-    return L2_reg * ((w1 ** 2).sum() + (w2 ** 2).sum())
-
-
-def update(eta, param, cost):
-    return (param, param - (eta * T.grad(cost, param)))
-
-
-def main(train_loc, eval_loc, model_dir):
-    learning_rate = 0.01
-    L1_reg = 0.00
-    L2_reg = 0.0001
-
-    print "... reading the data"
-    gold_train = list(read_json_file(train_loc))
-    print '... building the model'
-    pos_model_dir = path.join(model_dir, 'pos')
-    if path.exists(pos_model_dir):
-        shutil.rmtree(pos_model_dir)
-    os.mkdir(pos_model_dir)
-
-    setup_model_dir(sorted(POS_TAGS.keys()), POS_TAGS, POS_TEMPLATES, pos_model_dir)
-
-    train_model, evaluate_model = build_model(n_hidden, len(POS_TAGS), learning_rate,
-                                              L1_reg, L2_reg)
-
-    print '... training'
-    for epoch in range(1, n_epochs+1):
-        for raw_text, sents in gold_tuples:
-            for (ids, words, tags, ner, heads, deps), _ in sents:
-                tokens = nlp.tokenizer.tokens_from_list(words)
-                for t in tokens:
-                    guess = train_model([t.orth], [t.tag])
-                    loss += guess != t.tag
-        print loss
-        # compute zero-one loss on validation set
-        #error = numpy.mean([evaluate_model(x, y) for x, y in dev_examples])
-        #print('epoch %i, validation error %f %%' % (epoch, error * 100))
-
-
-if __name__ == '__main__':
-    plac.call(main)
diff --git a/spacy/_theano.pxd b/spacy/_theano.pxd
deleted file mode 100644
index cad0736c2..000000000
--- a/spacy/_theano.pxd
+++ /dev/null
@@ -1,13 +0,0 @@
-from ._ml cimport Model
-from thinc.nn cimport InputLayer
-
-
-cdef class TheanoModel(Model):
-    cdef InputLayer input_layer
-    cdef object train_func
-    cdef object predict_func
-    cdef object debug
-
-    cdef public float eta
-    cdef public float mu
-    cdef public float t
diff --git a/spacy/_theano.pyx b/spacy/_theano.pyx
deleted file mode 100644
index cc6886321..000000000
--- a/spacy/_theano.pyx
+++ /dev/null
@@ -1,52 +0,0 @@
-from thinc.api cimport Example, ExampleC
-from thinc.typedefs cimport weight_t
-
-from ._ml cimport arg_max_if_true
-from ._ml cimport arg_max_if_zero
-
-import numpy
-from os import path
-
-
-cdef class TheanoModel(Model):
-    def __init__(self, n_classes, input_spec, train_func, predict_func, model_loc=None,
-                 eta=0.001, mu=0.9, debug=None):
-        if model_loc is not None and path.isdir(model_loc):
-            model_loc = path.join(model_loc, 'model')
-
-        self.eta = eta
-        self.mu = mu
-        self.t = 1
-        initializer = lambda: 0.2 * numpy.random.uniform(-1.0, 1.0)
-        self.input_layer = InputLayer(input_spec, initializer)
-        self.train_func = train_func
-        self.predict_func = predict_func
-        self.debug = debug
-
-        self.n_classes = n_classes
-        self.n_feats = len(self.input_layer)
-        self.model_loc = model_loc
-        
-    def predict(self, Example eg):
-        self.input_layer.fill(eg.embeddings, eg.atoms, use_avg=True)
-        theano_scores = self.predict_func(eg.embeddings)[0]
-        cdef int i
-        for i in range(self.n_classes):
-            eg.c.scores[i] = theano_scores[i]
-        eg.c.guess = arg_max_if_true(eg.c.scores, eg.c.is_valid, self.n_classes)
-
-    def train(self, Example eg):
-        self.input_layer.fill(eg.embeddings, eg.atoms, use_avg=False)
-        theano_scores, update, y, loss = self.train_func(eg.embeddings, eg.costs,
-                                                         self.eta, self.mu)
-        self.input_layer.update(update, eg.atoms, self.t, self.eta, self.mu)
-        for i in range(self.n_classes):
-            eg.c.scores[i] = theano_scores[i]
-        eg.c.guess = arg_max_if_true(eg.c.scores, eg.c.is_valid, self.n_classes)
-        eg.c.best = arg_max_if_zero(eg.c.scores, eg.c.costs, self.n_classes)
-        eg.c.cost = eg.c.costs[eg.c.guess]
-        eg.c.loss = loss
-        self.t += 1
-
-    def end_training(self):
-        pass
diff --git a/spacy/de/__init__.py b/spacy/de/__init__.py
index 262fbe289..d7cc3dc65 100644
--- a/spacy/de/__init__.py
+++ b/spacy/de/__init__.py
@@ -6,6 +6,4 @@ from ..language import Language
 
 
 class German(Language):
-    @classmethod
-    def default_data_dir(cls):
-        return path.join(path.dirname(__file__), 'data')
+    pass
diff --git a/spacy/en/__init__.py b/spacy/en/__init__.py
index 4d057db20..4187a4890 100644
--- a/spacy/en/__init__.py
+++ b/spacy/en/__init__.py
@@ -4,8 +4,6 @@ from os import path
 
 from ..language import Language
 
-LOCAL_DATA_DIR = path.join(path.dirname(__file__), 'data')
-
 
 # improved list from Stone, Denis, Kwantes (2010)
 STOPWORDS = """
@@ -35,9 +33,7 @@ your yours yourself yourselves
 STOPWORDS = set(w for w in STOPWORDS.split() if w)
 
 class English(Language):
-    @classmethod
-    def default_data_dir(cls):
-        return LOCAL_DATA_DIR
+    lang = 'en'
 
     @staticmethod
     def is_stop(string):
diff --git a/spacy/en/download.py b/spacy/en/download.py
index 26d2b44be..f91b44601 100644
--- a/spacy/en/download.py
+++ b/spacy/en/download.py
@@ -8,8 +8,11 @@ from sputnik import Sputnik
 
 def migrate(path):
     data_path = os.path.join(path, 'data')
-    if os.path.isdir(data_path) and not os.path.islink(data_path):
-        shutil.rmtree(data_path)
+    if os.path.isdir(data_path):
+        if os.path.islink(data_path):
+            os.unlink(data_path)
+        else:
+            shutil.rmtree(data_path)
     for filename in os.listdir(path):
         if filename.endswith('.tgz'):
             os.unlink(os.path.join(path, filename))
@@ -17,8 +20,15 @@ def migrate(path):
 
 def link(package, path):
     if os.path.exists(path):
-        os.unlink(path)
-    os.symlink(package.dir_path('data'), path)
+        if os.path.isdir(path):
+            shutil.rmtree(path)
+        else:
+            os.unlink(path)
+
+    if not hasattr(os, 'symlink'):  # not supported by win+py27
+        shutil.copytree(package.dir_path('data'), path)
+    else:
+        os.symlink(package.dir_path('data'), path)
 
 
 @plac.annotations(
@@ -26,12 +36,16 @@ def link(package, path):
 )
 def main(data_size='all', force=False):
     # TODO read version from the same source as the setup
-    sputnik = Sputnik('spacy', '0.99.0', console=sys.stdout)
+    sputnik = Sputnik('spacy', '0.100.0', console=sys.stdout)
 
     path = os.path.dirname(os.path.abspath(__file__))
 
-    command = sputnik.make_command(
-        data_path=os.path.abspath(os.path.join(path, '..', 'data')),
+    data_path = os.path.abspath(os.path.join(path, '..', 'data'))
+    if not os.path.isdir(data_path):
+        os.mkdir(data_path)
+
+    command = sputnik.command(
+        data_path=data_path,
         repository_url='https://index.spacy.io')
 
     if force:
@@ -42,9 +56,6 @@ def main(data_size='all', force=False):
     # FIXME clean up old-style packages
     migrate(path)
 
-    # FIXME supply spacy with an old-style data dir
-    link(package, os.path.join(path, 'data'))
-
 
 if __name__ == '__main__':
     plac.call(main)
diff --git a/spacy/fi/__init__.py b/spacy/fi/__init__.py
index 8e7173767..adba064a5 100644
--- a/spacy/fi/__init__.py
+++ b/spacy/fi/__init__.py
@@ -6,6 +6,4 @@ from ..language import Language
 
 
 class Finnish(Language):
-    @classmethod
-    def default_data_dir(cls):
-        return path.join(path.dirname(__file__), 'data')
+    pass
diff --git a/spacy/it/__init__.py b/spacy/it/__init__.py
index a494de41f..6a824fe82 100644
--- a/spacy/it/__init__.py
+++ b/spacy/it/__init__.py
@@ -6,6 +6,4 @@ from ..language import Language
 
 
 class Italian(Language):
-    @classmethod
-    def default_data_dir(cls):
-        return path.join(path.dirname(__file__), 'data')
+    pass
diff --git a/spacy/language.py b/spacy/language.py
index f598518e2..fe03db4f5 100644
--- a/spacy/language.py
+++ b/spacy/language.py
@@ -20,9 +20,12 @@ from .syntax.ner import BiluoPushDown
 from .syntax.arc_eager import ArcEager
 
 from .attrs import TAG, DEP, ENT_IOB, ENT_TYPE, HEAD
+from .util import get_package
 
 
 class Language(object):
+    lang = None
+
     @staticmethod
     def lower(string):
         return string.lower()
@@ -100,7 +103,7 @@ class Language(object):
         return 0
 
     @classmethod
-    def default_lex_attrs(cls, data_dir=None):
+    def default_lex_attrs(cls):
         return {
             attrs.LOWER: cls.lower,
             attrs.NORM: cls.norm,
@@ -134,85 +137,108 @@ class Language(object):
         return {0: {'PER': True, 'LOC': True, 'ORG': True, 'MISC': True}}
 
     @classmethod
-    def default_data_dir(cls):
-        return path.join(path.dirname(__file__), 'data')
-
-    @classmethod
-    def default_vocab(cls, data_dir=None, get_lex_attr=None):
-        if data_dir is None:
-            data_dir = cls.default_data_dir()
+    def default_vocab(cls, package=None, get_lex_attr=None):
+        if package is None:
+            package = get_package()
         if get_lex_attr is None:
-            get_lex_attr = cls.default_lex_attrs(data_dir)
-        return Vocab.from_dir(
-                path.join(data_dir, 'vocab'),
-                get_lex_attr=get_lex_attr)
+            get_lex_attr = cls.default_lex_attrs()
+        return Vocab.from_package(package, get_lex_attr=get_lex_attr)
 
     @classmethod
-    def default_tokenizer(cls, vocab, data_dir):
-        if path.exists(data_dir):
-            return Tokenizer.from_dir(vocab, data_dir)
-        else:
-            return Tokenizer(vocab, {}, None, None, None)
-
-    @classmethod
-    def default_tagger(cls, vocab, data_dir):
-        if path.exists(data_dir):
-            return Tagger.from_dir(data_dir, vocab)
-        else:
-            return None
-
-    @classmethod
-    def default_parser(cls, vocab, data_dir):
-        if path.exists(data_dir):
+    def default_parser(cls, package, vocab):
+        data_dir = package.dir_path('deps', require=False)
+        if data_dir and path.exists(data_dir):
             return Parser.from_dir(data_dir, vocab.strings, ArcEager)
-        else:
-            return None
 
     @classmethod
-    def default_entity(cls, vocab, data_dir):
-        if path.exists(data_dir):
+    def default_entity(cls, package, vocab):
+        data_dir = package.dir_path('ner', require=False)
+        if data_dir and path.exists(data_dir):
             return Parser.from_dir(data_dir, vocab.strings, BiluoPushDown)
-        else:
-            return None
 
-    @classmethod
-    def default_matcher(cls, vocab, data_dir):
-        if path.exists(data_dir):
-            return Matcher.from_dir(data_dir, vocab)
-        else:
-            return None
+    def __init__(self,
+        data_dir=None,
+        model=None,
+        vocab=None,
+        tokenizer=None,
+        tagger=None,
+        parser=None,
+        entity=None,
+        matcher=None,
+        serializer=None,
+        load_vectors=True):
+        """
+           a model can be specified:
 
-    def __init__(self, data_dir=None, vocab=None, tokenizer=None, tagger=None,
-                 parser=None, entity=None, matcher=None, serializer=None,
-                 load_vectors=True):
+           1) by a path to the model directory (DEPRECATED)
+             - Language(data_dir='path/to/data')
+
+           2) by a language identifier (and optionally a package root dir)
+             - Language(lang='en')
+             - Language(lang='en', data_dir='spacy/data')
+
+           3) by a model name/version (and optionally a package root dir)
+             - Language(model='en_default')
+             - Language(model='en_default ==1.0.0')
+             - Language(model='en_default <1.1.0, data_dir='spacy/data')
+        """
+        # support non-package data dirs
+        if data_dir and path.exists(path.join(data_dir, 'vocab')):
+            class Package(object):
+                def __init__(self, root):
+                    self.root = root
+
+                def has_file(self, *path_parts):
+                    return path.exists(path.join(self.root, *path_parts))
+
+                def file_path(self, *path_parts, **kwargs):
+                    return path.join(self.root, *path_parts)
+
+                def dir_path(self, *path_parts, **kwargs):
+                    return path.join(self.root, *path_parts)
+
+                def load_utf8(self, func, *path_parts, **kwargs):
+                    with io.open(self.file_path(path.join(*path_parts)),
+                                 mode='r', encoding='utf8') as f:
+                        return func(f)
+
+            warn("using non-package data_dir", DeprecationWarning)
+            package = Package(data_dir)
+        else:
+            package = get_package(name=model, data_path=data_dir)
         if load_vectors is not True:
             warn("load_vectors is deprecated", DeprecationWarning)
-        if data_dir in (None, True):
-            data_dir = self.default_data_dir()
         if vocab in (None, True):
-            vocab = self.default_vocab(data_dir)
-        if tokenizer in (None, True):
-            tokenizer = self.default_tokenizer(vocab, data_dir=path.join(data_dir, 'tokenizer'))
-        if tagger in (None, True):
-            tagger = self.default_tagger(vocab, data_dir=path.join(data_dir, 'pos'))
-        if entity in (None, True):
-            entity = self.default_entity(vocab, data_dir=path.join(data_dir, 'ner'))
-        if parser in (None, True):
-            parser = self.default_parser(vocab, data_dir=path.join(data_dir, 'deps'))
-        if matcher in (None, True):
-            matcher = self.default_matcher(vocab, data_dir=data_dir)
+            self.vocab = self.default_vocab(package)
         self.vocab = vocab
+        if tokenizer in (None, True):
+            self.tokenizer = Tokenizer.from_package(package, self.vocab)
         self.tokenizer = tokenizer
+        if tagger in (None, True):
+            tagger = Tagger.from_package(package, self.vocab)
         self.tagger = tagger
-        self.parser = parser
+        if entity in (None, True):
+            entity = self.default_entity(package, self.vocab)
         self.entity = entity
+        if parser in (None, True):
+            parser = self.default_parser(package, self.vocab)
+        self.parser = parser
+        if matcher in (None, True):
+            matcher = Matcher.from_package(package, self.vocab)
         self.matcher = matcher
 
     def __reduce__(self):
-        return (self.__class__,
-                  (None, self.vocab, self.tokenizer, self.tagger, self.parser,
-                   self.entity, self.matcher, None),
-                None, None)
+        args = (
+            None, # data_dir
+            None, # model
+            self.vocab,
+            self.tokenizer,
+            self.tagger,
+            self.parser,
+            self.entity,
+            self.matcher
+        )
+        return (self.__class__, args, None, None)
 
     def __call__(self, text, tag=True, parse=True, entity=True):
         """Apply the pipeline to some text.  The text can span multiple sentences,
diff --git a/spacy/lemmatizer.py b/spacy/lemmatizer.py
index 08e511f68..c5b9c1c50 100644
--- a/spacy/lemmatizer.py
+++ b/spacy/lemmatizer.py
@@ -12,16 +12,21 @@ from .parts_of_speech import NOUN, VERB, ADJ, PUNCT
 
 class Lemmatizer(object):
     @classmethod
-    def from_dir(cls, data_dir):
+    def from_package(cls, package):
         index = {}
         exc = {}
         for pos in ['adj', 'noun', 'verb']:
-            index[pos] = read_index(path.join(data_dir, 'wordnet', 'index.%s' % pos))
-            exc[pos] = read_exc(path.join(data_dir, 'wordnet', '%s.exc' % pos))
-        if path.exists(path.join(data_dir, 'vocab', 'lemma_rules.json')):
-            rules = json.load(codecs.open(path.join(data_dir, 'vocab', 'lemma_rules.json'), encoding='utf_8'))
-        else:
-            rules = {}
+            index[pos] = package.load_utf8(read_index,
+                'wordnet', 'index.%s' % pos,
+                default=set())  # TODO: really optional?
+            exc[pos] = package.load_utf8(read_exc,
+                'wordnet', '%s.exc' % pos,
+                default={})  # TODO: really optional?
+
+        rules = package.load_utf8(json.load,
+            'vocab', 'lemma_rules.json',
+            default={})  # TODO: really optional?
+
         return cls(index, exc, rules)
 
     def __init__(self, index, exceptions, rules):
@@ -70,11 +75,9 @@ def lemmatize(string, index, exceptions, rules):
     return set(forms)
 
 
-def read_index(loc):
+def read_index(fileobj):
     index = set()
-    if not path.exists(loc):
-        return index
-    for line in codecs.open(loc, 'r', 'utf8'):
+    for line in fileobj:
         if line.startswith(' '):
             continue
         pieces = line.split()
@@ -84,11 +87,9 @@ def read_index(loc):
     return index
 
 
-def read_exc(loc):
+def read_exc(fileobj):
     exceptions = {}
-    if not path.exists(loc):
-        return exceptions
-    for line in codecs.open(loc, 'r', 'utf8'):
+    for line in fileobj:
         if line.startswith(' '):
             continue
         pieces = line.split()
diff --git a/spacy/matcher.pyx b/spacy/matcher.pyx
index 1fa91fab1..4d36b7742 100644
--- a/spacy/matcher.pyx
+++ b/spacy/matcher.pyx
@@ -169,14 +169,11 @@ cdef class Matcher:
     cdef object _patterns
 
     @classmethod
-    def from_dir(cls, data_dir, Vocab vocab):
-        patterns_loc = path.join(data_dir, 'vocab', 'gazetteer.json')
-        if path.exists(patterns_loc):
-            patterns_data = open(patterns_loc).read()
-            patterns = json.loads(patterns_data)
-            return cls(vocab, patterns)
-        else:
-            return cls(vocab, {})
+    def from_package(cls, package, Vocab vocab):
+        patterns = package.load_utf8(json.load,
+            'vocab', 'gazetteer.json',
+            default={})  # TODO: really optional?
+        return cls(vocab, patterns)
 
     def __init__(self, vocab, patterns):
         self.vocab = vocab
diff --git a/spacy/senses.pxd b/spacy/senses.pxd
deleted file mode 100644
index 5ebe89660..000000000
--- a/spacy/senses.pxd
+++ /dev/null
@@ -1,62 +0,0 @@
-# Enum of Wordnet supersenses
-cimport parts_of_speech
-from .typedefs cimport flags_t
-
-cpdef enum:
-    A_behavior
-    A_body
-    A_feeling
-    A_mind
-    A_motion
-    A_perception
-    A_quantity
-    A_relation
-    A_social
-    A_spatial
-    A_substance
-    A_time
-    A_weather
-    N_act
-    N_animal
-    N_artifact
-    N_attribute
-    N_body
-    N_cognition
-    N_communication
-    N_event
-    N_feeling
-    N_food
-    N_group
-    N_location
-    N_motive
-    N_object
-    N_person
-    N_phenomenon
-    N_plant
-    N_possession
-    N_process
-    N_quantity
-    N_relation
-    N_shape
-    N_state
-    N_substance
-    N_time
-    V_body
-    V_change
-    V_cognition
-    V_communication
-    V_competition
-    V_consumption
-    V_contact
-    V_creation
-    V_emotion
-    V_motion
-    V_perception
-    V_possession
-    V_social
-    V_stative
-    V_weather
-
-
-cdef flags_t[<int>parts_of_speech.N_UNIV_TAGS] POS_SENSES
-
diff --git a/spacy/senses.pyx b/spacy/senses.pyx
deleted file mode 100644
index 99d193774..000000000
--- a/spacy/senses.pyx
+++ /dev/null
@@ -1,88 +0,0 @@
-from __future__ import unicode_literals
-cimport parts_of_speech
-
-
-POS_SENSES[<int>parts_of_speech.NO_TAG] = 0
-POS_SENSES[<int>parts_of_speech.ADJ] = 0
-POS_SENSES[<int>parts_of_speech.ADV] = 0
-POS_SENSES[<int>parts_of_speech.ADP] = 0
-POS_SENSES[<int>parts_of_speech.CONJ] = 0
-POS_SENSES[<int>parts_of_speech.DET] = 0
-POS_SENSES[<int>parts_of_speech.NOUN] = 0
-POS_SENSES[<int>parts_of_speech.NUM] = 0
-POS_SENSES[<int>parts_of_speech.PRON] = 0
-POS_SENSES[<int>parts_of_speech.PRT] = 0
-POS_SENSES[<int>parts_of_speech.VERB] = 0
-POS_SENSES[<int>parts_of_speech.X] = 0
-POS_SENSES[<int>parts_of_speech.PUNCT] = 0
-POS_SENSES[<int>parts_of_speech.EOL] = 0
-
-
-cdef int _sense = 0
-
-for _sense in range(A_behavior, N_act):
-    POS_SENSES[<int>parts_of_speech.ADJ] |= 1 << _sense
-
-for _sense in range(N_act, V_body):
-    POS_SENSES[<int>parts_of_speech.NOUN] |= 1 << _sense
-
-for _sense in range(V_body, V_weather+1):
-    POS_SENSES[<int>parts_of_speech.VERB] |= 1 << _sense
-
-
-
-STRINGS = (
-    'A_behavior',
-    'A_body',
-    'A_feeling',
-    'A_mind',
-    'A_motion',
-    'A_perception',
-    'A_quantity',
-    'A_relation',
-    'A_social',
-    'A_spatial',
-    'A_substance',
-    'A_time',
-    'A_weather',
-    'N_act',
-    'N_animal',
-    'N_artifact',
-    'N_attribute',
-    'N_body',
-    'N_cognition',
-    'N_communication',
-    'N_event',
-    'N_feeling',
-    'N_food',
-    'N_group',
-    'N_location',
-    'N_motive',
-    'N_object',
-    'N_person',
-    'N_phenomenon',
-    'N_plant',
-    'N_possession',
-    'N_process',
-    'N_quantity',
-    'N_relation',
-    'N_shape',
-    'N_state',
-    'N_substance',
-    'N_time',
-    'V_body',
-    'V_change',
-    'V_cognition',
-    'V_communication',
-    'V_competition',
-    'V_consumption',
-    'V_contact',
-    'V_creation',
-    'V_emotion',
-    'V_motion',
-    'V_perception',
-    'V_possession',
-    'V_social',
-    'V_stative',
-    'V_weather'
-)
diff --git a/spacy/tagger.pyx b/spacy/tagger.pyx
index 421d97357..2c05b4a84 100644
--- a/spacy/tagger.pyx
+++ b/spacy/tagger.pyx
@@ -146,15 +146,19 @@ cdef class Tagger:
         return cls(vocab, model)
 
     @classmethod
-    def from_dir(cls, data_dir, vocab):
-        if path.exists(path.join(data_dir, 'templates.json')):
-            templates = json.loads(open(path.join(data_dir, 'templates.json')))
-        else:
-            templates = cls.default_templates()
+    def from_package(cls, package, vocab):
+        # TODO: templates.json deprecated? not present in latest package
+        templates = cls.default_templates()
+        # templates = package.load_utf8(json.load,
+        #     'pos', 'templates.json',
+        #     default=cls.default_templates())
+
         model = TaggerModel(vocab.morphology.n_tags,
             ConjunctionExtracter(N_CONTEXT_FIELDS, templates))
-        if path.exists(path.join(data_dir, 'model')):
-            model.load(path.join(data_dir, 'model'))
+
+        if package.has_file('pos', 'model'):  # TODO: really optional?
+            model.load(package.file_path('pos', 'model'))
+
         return cls(vocab, model)
 
     def __init__(self, Vocab vocab, TaggerModel model):
diff --git a/spacy/tests/conftest.py b/spacy/tests/conftest.py
index 6aee2c31e..03e728a12 100644
--- a/spacy/tests/conftest.py
+++ b/spacy/tests/conftest.py
@@ -1,12 +1,11 @@
+from spacy.en import English
+
 import pytest
-from spacy.en import English, LOCAL_DATA_DIR
-import os
 
 
 @pytest.fixture(scope="session")
 def EN():
-    data_dir = os.environ.get('SPACY_DATA', LOCAL_DATA_DIR)
-    return English(data_dir=data_dir)
+    return English()
 
 
 def pytest_addoption(parser):
diff --git a/spacy/tests/de/test_de.py b/spacy/tests/de/test_de.py
deleted file mode 100644
index e0b941c36..000000000
--- a/spacy/tests/de/test_de.py
+++ /dev/null
@@ -1,12 +0,0 @@
-# encoding: utf8
-from __future__ import unicode_literals
-
-import spacy.de
-
-
-#def test_tokenizer():
-#    lang = spacy.de.German()
-#
-#    doc = lang(u'Biografie: Ein Spiel ist ein Theaterstück des Schweizer Schriftstellers Max Frisch, das 1967 entstand und am 1. Februar 1968 im Schauspielhaus Zürich uraufgeführt wurde. 1984 legte Frisch eine überarbeitete Neufassung vor. Das von Frisch als Komödie bezeichnete Stück greift eines seiner zentralen Themen auf: die Möglichkeit oder Unmöglichkeit des Menschen, seine Identität zu verändern.')
-#    for token in doc:
-#        print(repr(token.string))
diff --git a/spacy/tests/serialize/test_io.py b/spacy/tests/serialize/test_io.py
index f90bb20c2..3e269873f 100644
--- a/spacy/tests/serialize/test_io.py
+++ b/spacy/tests/serialize/test_io.py
@@ -4,6 +4,9 @@ from spacy.serialize.packer import Packer
 from spacy.attrs import ORTH, SPACY
 from spacy.tokens import Doc
 import math
+import tempfile
+import shutil
+import os
 
 
 @pytest.mark.models
@@ -11,17 +14,21 @@ def test_read_write(EN):
     doc1 = EN(u'This is a simple test. With a couple of sentences.')
     doc2 = EN(u'This is another test document.')
 
-    with open('/tmp/spacy_docs.bin', 'wb') as file_:
-        file_.write(doc1.to_bytes())
-        file_.write(doc2.to_bytes())
+    try:
+        tmp_dir = tempfile.mkdtemp()
+        with open(os.path.join(tmp_dir, 'spacy_docs.bin'), 'wb') as file_:
+            file_.write(doc1.to_bytes())
+            file_.write(doc2.to_bytes())
 
-    with open('/tmp/spacy_docs.bin', 'rb') as file_:
-        bytes1, bytes2 = Doc.read_bytes(file_)
-        r1 = Doc(EN.vocab).from_bytes(bytes1)
-        r2 = Doc(EN.vocab).from_bytes(bytes2)
+        with open(os.path.join(tmp_dir, 'spacy_docs.bin'), 'rb') as file_:
+            bytes1, bytes2 = Doc.read_bytes(file_)
+            r1 = Doc(EN.vocab).from_bytes(bytes1)
+            r2 = Doc(EN.vocab).from_bytes(bytes2)
 
-    assert r1.string == doc1.string
-    assert r2.string == doc2.string
+        assert r1.string == doc1.string
+        assert r2.string == doc2.string
+    finally:
+        shutil.rmtree(tmp_dir)
 
 
 @pytest.mark.models
diff --git a/spacy/tests/serialize/test_packer.py b/spacy/tests/serialize/test_packer.py
index e0d24208a..0e13b2de5 100644
--- a/spacy/tests/serialize/test_packer.py
+++ b/spacy/tests/serialize/test_packer.py
@@ -10,7 +10,6 @@ from spacy.en import English
 from spacy.vocab import Vocab
 from spacy.tokens.doc import Doc
 from spacy.tokenizer import Tokenizer
-from spacy.en import LOCAL_DATA_DIR
 from os import path
 
 from spacy.attrs import ORTH, SPACY, TAG, DEP, HEAD
diff --git a/spacy/tests/spans/conftest.py b/spacy/tests/spans/conftest.py
index f63816e05..d8a0a4cb9 100644
--- a/spacy/tests/spans/conftest.py
+++ b/spacy/tests/spans/conftest.py
@@ -1,9 +1,8 @@
 import pytest
-from spacy.en import English, LOCAL_DATA_DIR
+from spacy.en import English
 import os
 
 
 @pytest.fixture(scope="session")
 def en_nlp():
-    data_dir = os.environ.get('SPACY_DATA', LOCAL_DATA_DIR)
-    return English(data_dir=data_dir)
+    return English()
diff --git a/spacy/tests/tagger/test_lemmatizer.py b/spacy/tests/tagger/test_lemmatizer.py
index 91aa8ee65..e25fbe199 100644
--- a/spacy/tests/tagger/test_lemmatizer.py
+++ b/spacy/tests/tagger/test_lemmatizer.py
@@ -4,31 +4,33 @@ import io
 import pickle
 
 from spacy.lemmatizer import Lemmatizer, read_index, read_exc
-from spacy.en import LOCAL_DATA_DIR
-from os import path
+from spacy.util import get_package
 
 import pytest
 
 
-def test_read_index():
-    wn = path.join(LOCAL_DATA_DIR, 'wordnet')
-    index = read_index(path.join(wn, 'index.noun'))
+@pytest.fixture
+def package():
+    return get_package()
+
+
+@pytest.fixture
+def lemmatizer(package):
+    return Lemmatizer.from_package(package)
+
+
+def test_read_index(package):
+    index = package.load_utf8(read_index, 'wordnet', 'index.noun')
     assert 'man' in index
     assert 'plantes' not in index
     assert 'plant' in index
 
 
-def test_read_exc():
-    wn = path.join(LOCAL_DATA_DIR, 'wordnet')
-    exc = read_exc(path.join(wn, 'verb.exc'))
+def test_read_exc(package):
+    exc = package.load_utf8(read_exc, 'wordnet', 'verb.exc')
     assert exc['was'] == ('be',)
 
 
-@pytest.fixture
-def lemmatizer():
-    return Lemmatizer.from_dir(path.join(LOCAL_DATA_DIR))
-
-
 def test_noun_lemmas(lemmatizer):
     do = lemmatizer.noun
 
diff --git a/spacy/tests/tokens/test_token_references.py b/spacy/tests/tokens/test_token_references.py
index e8dbff6fe..24639f141 100644
--- a/spacy/tests/tokens/test_token_references.py
+++ b/spacy/tests/tokens/test_token_references.py
@@ -2,16 +2,15 @@ from __future__ import unicode_literals
 import pytest
 import gc
 
-from spacy.en import English, LOCAL_DATA_DIR
+from spacy.en import English
 import os
 
-data_dir = os.environ.get('SPACY_DATA', LOCAL_DATA_DIR)
 # Let this have its own instances, as we have to be careful about memory here
 # that's the point, after all
 
 @pytest.mark.models
 def get_orphan_token(text, i):
-    nlp = English(data_dir=data_dir)
+    nlp = English()
     tokens = nlp(text)
     gc.collect()
     token = tokens[i]
@@ -41,7 +40,7 @@ def _orphan_from_list(toks):
 @pytest.mark.models
 def test_list_orphans():
     # Test case from NSchrading
-    nlp = English(data_dir=data_dir)
+    nlp = English()
     samples = ["a", "test blah wat okay"]
     lst = []
     for sample in samples:
diff --git a/spacy/tests/website/conftest.py b/spacy/tests/website/conftest.py
index 35c38d845..b4934d20b 100644
--- a/spacy/tests/website/conftest.py
+++ b/spacy/tests/website/conftest.py
@@ -5,9 +5,8 @@ import os
 
 @pytest.fixture(scope='session')
 def nlp():
-    from spacy.en import English, LOCAL_DATA_DIR
-    data_dir = os.environ.get('SPACY_DATA', LOCAL_DATA_DIR)
-    return English(data_dir=data_dir)
+    from spacy.en import English
+    return English()
 
 
 @pytest.fixture()
diff --git a/spacy/tests/website/test_api.py b/spacy/tests/website/test_api.py
index 6eb1d1a03..39eb38dd6 100644
--- a/spacy/tests/website/test_api.py
+++ b/spacy/tests/website/test_api.py
@@ -75,7 +75,7 @@ def test_count_by(nlp):
 @pytest.mark.models
 def test_read_bytes(nlp):
     from spacy.tokens.doc import Doc
-    loc = '/tmp/test_serialize.bin'
+    loc = 'test_serialize.bin'
     with open(loc, 'wb') as file_:
         file_.write(nlp(u'This is a document.').to_bytes())
         file_.write(nlp(u'This is another.').to_bytes())
diff --git a/spacy/tests/website/test_home.py b/spacy/tests/website/test_home.py
index d03acf855..5317e7e56 100644
--- a/spacy/tests/website/test_home.py
+++ b/spacy/tests/website/test_home.py
@@ -10,9 +10,8 @@ def token(doc):
 
 
 def test_load_resources_and_process_text():
-    from spacy.en import English, LOCAL_DATA_DIR
-    data_dir = os.environ.get('SPACY_DATA', LOCAL_DATA_DIR)
-    nlp = English(data_dir=data_dir)
+    from spacy.en import English
+    nlp = English()
     doc = nlp('Hello, world. Here are two sentences.')
 
 
@@ -154,9 +153,9 @@ def test_efficient_binary_serialization(doc):
     from spacy.tokens.doc import Doc
 
     byte_string = doc.to_bytes()
-    open('/tmp/moby_dick.bin', 'wb').write(byte_string)
+    open('moby_dick.bin', 'wb').write(byte_string)
 
     nlp = spacy.en.English()
-    for byte_string in Doc.read_bytes(open('/tmp/moby_dick.bin', 'rb')):
+    for byte_string in Doc.read_bytes(open('moby_dick.bin', 'rb')):
        doc = Doc(nlp.vocab)
        doc.from_bytes(byte_string)
diff --git a/spacy/tokenizer.pyx b/spacy/tokenizer.pyx
index 1cde1e76e..345734682 100644
--- a/spacy/tokenizer.pyx
+++ b/spacy/tokenizer.pyx
@@ -41,8 +41,8 @@ cdef class Tokenizer:
         return (self.__class__, args, None, None)
 
     @classmethod
-    def from_dir(cls, Vocab vocab, data_dir):
-        rules, prefix_re, suffix_re, infix_re = read_lang_data(data_dir)
+    def from_package(cls, package, Vocab vocab):
+        rules, prefix_re, suffix_re, infix_re = read_lang_data(package)
         prefix_re = re.compile(prefix_re)
         suffix_re = re.compile(suffix_re)
         infix_re = re.compile(infix_re)
diff --git a/spacy/util.py b/spacy/util.py
index 849a3e219..8c9ea319c 100644
--- a/spacy/util.py
+++ b/spacy/util.py
@@ -1,10 +1,24 @@
-from os import path
+import os
 import io
 import json
 import re
+
+from sputnik import Sputnik
+
 from .attrs import TAG, HEAD, DEP, ENT_IOB, ENT_TYPE
 
-DATA_DIR = path.join(path.dirname(__file__), '..', 'data')
+
+def get_package(name=None, data_path=None):
+    if data_path is None:
+        if os.environ.get('SPACY_DATA'):
+            data_path = os.environ.get('SPACY_DATA')
+        else:
+            data_path = os.path.abspath(
+                os.path.join(os.path.dirname(__file__), 'data'))
+
+    sputnik = Sputnik('spacy', '0.100.0')  # TODO: retrieve version
+    pool = sputnik.pool(data_path)
+    return pool.get(name or 'en_default')
 
 
 def normalize_slice(length, start, stop, step=None):
@@ -31,67 +45,63 @@ def utf8open(loc, mode='r'):
     return io.open(loc, mode, encoding='utf8')
 
 
-def read_lang_data(data_dir):
-    with open(path.join(data_dir, 'specials.json')) as file_:
-        tokenization = json.load(file_)
-    prefix = read_prefix(data_dir)
-    suffix = read_suffix(data_dir)
-    infix = read_infix(data_dir)
+def read_lang_data(package):
+    tokenization = package.load_utf8(json.load, 'tokenizer', 'specials.json')
+    prefix = package.load_utf8(read_prefix, 'tokenizer', 'prefix.txt')
+    suffix = package.load_utf8(read_suffix, 'tokenizer', 'suffix.txt')
+    infix = package.load_utf8(read_infix, 'tokenizer', 'infix.txt')
     return tokenization, prefix, suffix, infix
 
 
-def read_prefix(data_dir):
-    with  utf8open(path.join(data_dir, 'prefix.txt')) as file_:
-        entries = file_.read().split('\n')
-        expression = '|'.join(['^' + re.escape(piece) for piece in entries if piece.strip()])
+def read_prefix(fileobj):
+    entries = fileobj.read().split('\n')
+    expression = '|'.join(['^' + re.escape(piece) for piece in entries if piece.strip()])
     return expression
 
 
-def read_suffix(data_dir):
-    with utf8open(path.join(data_dir, 'suffix.txt')) as file_:
-        entries = file_.read().split('\n')
-        expression = '|'.join([piece + '$' for piece in entries if piece.strip()])
+def read_suffix(fileobj):
+    entries = fileobj.read().split('\n')
+    expression = '|'.join([piece + '$' for piece in entries if piece.strip()])
     return expression
 
 
-def read_infix(data_dir):
-    with utf8open(path.join(data_dir, 'infix.txt')) as file_:
-        entries = file_.read().split('\n')
-        expression = '|'.join([piece for piece in entries if piece.strip()])
+def read_infix(fileobj):
+    entries = fileobj.read().split('\n')
+    expression = '|'.join([piece for piece in entries if piece.strip()])
     return expression
 
 
-def read_tokenization(lang):
-    loc = path.join(DATA_DIR, lang, 'tokenization')
-    entries = []
-    seen = set()
-    with utf8open(loc) as file_:
-        for line in file_:
-            line = line.strip()
-            if line.startswith('#'):
-                continue
-            if not line:
-                continue
-            pieces = line.split()
-            chunk = pieces.pop(0)
-            assert chunk not in seen, chunk
-            seen.add(chunk)
-            entries.append((chunk, list(pieces)))
-            if chunk[0].isalpha() and chunk[0].islower():
-                chunk = chunk[0].title() + chunk[1:]
-                pieces[0] = pieces[0][0].title() + pieces[0][1:]
-                seen.add(chunk)
-                entries.append((chunk, pieces))
-    return entries
+# def read_tokenization(lang):
+#     loc = path.join(DATA_DIR, lang, 'tokenization')
+#     entries = []
+#     seen = set()
+#     with utf8open(loc) as file_:
+#         for line in file_:
+#             line = line.strip()
+#             if line.startswith('#'):
+#                 continue
+#             if not line:
+#                 continue
+#             pieces = line.split()
+#             chunk = pieces.pop(0)
+#             assert chunk not in seen, chunk
+#             seen.add(chunk)
+#             entries.append((chunk, list(pieces)))
+#             if chunk[0].isalpha() and chunk[0].islower():
+#                 chunk = chunk[0].title() + chunk[1:]
+#                 pieces[0] = pieces[0][0].title() + pieces[0][1:]
+#                 seen.add(chunk)
+#                 entries.append((chunk, pieces))
+#     return entries
 
 
-def read_detoken_rules(lang): # Deprecated?
-    loc = path.join(DATA_DIR, lang, 'detokenize')
-    entries = []
-    with utf8open(loc) as file_:
-        for line in file_:
-            entries.append(line.strip())
-    return entries
+# def read_detoken_rules(lang): # Deprecated?
+#     loc = path.join(DATA_DIR, lang, 'detokenize')
+#     entries = []
+#     with utf8open(loc) as file_:
+#         for line in file_:
+#             entries.append(line.strip())
+#     return entries
 
 
 def align_tokens(ref, indices): # Deprecated, surely?
diff --git a/spacy/vocab.pyx b/spacy/vocab.pyx
index 3817e7127..bb0ae6173 100644
--- a/spacy/vocab.pyx
+++ b/spacy/vocab.pyx
@@ -47,28 +47,27 @@ cdef class Vocab:
     '''A map container for a language's LexemeC structs.
     '''
     @classmethod
-    def from_dir(cls, data_dir, get_lex_attr=None):
-        if not path.exists(data_dir):
-            raise IOError("Directory %s not found -- cannot load Vocab." % data_dir)
-        if not path.isdir(data_dir):
-            raise IOError("Path %s is a file, not a dir -- cannot load Vocab." % data_dir)
+    def from_package(cls, package, get_lex_attr=None):
+        tag_map = package.load_utf8(json.load,
+            'vocab', 'tag_map.json')
+
+        lemmatizer = Lemmatizer.from_package(package)
+
+        serializer_freqs = package.load_utf8(json.load,
+            'vocab', 'serializer.json',
+            require=False)  # TODO: really optional?
 
-        tag_map = json.load(open(path.join(data_dir, 'tag_map.json')))
-        lemmatizer = Lemmatizer.from_dir(path.join(data_dir, '..'))
-        if path.exists(path.join(data_dir, 'serializer.json')):
-            serializer_freqs = json.load(open(path.join(data_dir, 'serializer.json')))
-        else:
-            serializer_freqs = None
         cdef Vocab self = cls(get_lex_attr=get_lex_attr, tag_map=tag_map,
                               lemmatizer=lemmatizer, serializer_freqs=serializer_freqs)
 
-        if path.exists(path.join(data_dir, 'strings.json')):
-            with io.open(path.join(data_dir, 'strings.json'), 'r', encoding='utf8') as file_:
-                self.strings.load(file_)
-            self.load_lexemes(path.join(data_dir, 'lexemes.bin'))
-                
-        if path.exists(path.join(data_dir, 'vec.bin')):
-            self.vectors_length = self.load_vectors_from_bin_loc(path.join(data_dir, 'vec.bin'))
+        if package.has_file('vocab', 'strings.json'):  # TODO: really optional?
+            package.load_utf8(self.strings.load, 'vocab', 'strings.json')
+            self.load_lexemes(package.file_path('vocab', 'lexemes.bin'))
+
+        if package.has_file('vocab', 'vec.bin'):  # TODO: really optional?
+            self.vectors_length = self.load_vectors_from_bin_loc(
+                package.file_path('vocab', 'vec.bin'))
+
         return self
 
     def __init__(self, get_lex_attr=None, tag_map=None, lemmatizer=None, serializer_freqs=None):
diff --git a/tox.ini b/tox.ini
new file mode 100644
index 000000000..a4b67ef54
--- /dev/null
+++ b/tox.ini
@@ -0,0 +1,13 @@
+[tox]
+envlist =
+    py27
+    py34
+recreate = True
+
+[testenv]
+changedir = {envtmpdir}
+deps =
+    pytest
+commands =
+    python -m spacy.en.download
+    python -m pytest {toxinidir}/spacy/ -x --models --vectors --slow
diff --git a/venv.ps1 b/venv.ps1
new file mode 100644
index 000000000..4fd5c157c
--- /dev/null
+++ b/venv.ps1
@@ -0,0 +1,32 @@
+param (
+    [string]$python = $(throw "-python is required."),
+    [string]$install_mode = $(throw "-install_mode is required."),
+    [string]$pip_date,
+    [string]$compiler
+)
+
+$ErrorActionPreference = "Stop"
+
+if(!(Test-Path -Path ".build"))
+{
+    if($compiler -eq "mingw32")
+    {
+        virtualenv .build --system-site-packages --python $python
+    }
+    else
+    {
+        virtualenv .build --python $python
+    }
+
+    if($compiler)
+    {
+        "[build]`r`ncompiler=$compiler" | Out-File -Encoding ascii .\.build\Lib\distutils\distutils.cfg
+    }
+}
+
+.build\Scripts\activate.ps1
+
+python build.py prepare $pip_date
+python build.py $install_mode
+python build.py test
+exit $LASTEXITCODE
diff --git a/venv.sh b/venv.sh
new file mode 100755
index 000000000..8852a5ba0
--- /dev/null
+++ b/venv.sh
@@ -0,0 +1,16 @@
+#!/bin/bash
+set -e
+
+if [ ! -d ".build" ]; then
+    virtualenv .build --python $1
+fi
+
+if [ -d ".build/bin" ]; then
+    source .build/bin/activate
+elif [ -d ".build/Scripts" ]; then
+    source .build/Scripts/activate
+fi
+
+python build.py prepare $3
+python build.py $2
+python build.py test
diff --git a/website/src/jade/404.jade b/website/src/jade/404.jade
index 9ed5366fa..cd9d5d96f 100644
--- a/website/src/jade/404.jade
+++ b/website/src/jade/404.jade
@@ -2,6 +2,7 @@ include ./header
 include ./mixins.jade
 
 - var Page = InitPage(Site, Authors.spacy, "home", '404')
+- Page.canonical_url = null
 - Page.is_error = true
 - Site.slogan = "404"
 - Page.active = {}
diff --git a/website/src/jade/blog/dead-code-should-be-buried/index.jade b/website/src/jade/blog/dead-code-should-be-buried/index.jade
index d5a778a54..af85cc140 100644
--- a/website/src/jade/blog/dead-code-should-be-buried/index.jade
+++ b/website/src/jade/blog/dead-code-should-be-buried/index.jade
@@ -4,7 +4,7 @@ include ./meta.jade
 
 +WritePost(Meta)
     section.intro
-        p Natural Language Processing moves fast, so maintaining a good library means constantly throwing things away.  Most libraries are failing badly at this, as academics hate to editorialize. This post explains the problem, why it's so damaging, and why I wrote #[a(href="http://spacy.io") spaCy] to do things differently.
+        p Natural Language Processing moves fast, so maintaining a good library means constantly throwing things away.  Most libraries are failing badly at this, as academics hate to editorialize. This post explains the problem, why it's so damaging, and why I wrote #[a(href="https://spacy.io") spaCy] to do things differently.
 
     p Imagine: you try to use Google Translate, but it asks you to first select which model you want. The new, awesome deep-learning model is there, but so are lots of others. You pick one that sounds fancy, but it turns out it's a 20-year old experimental model trained on a corpus of oven manuals. When it performs little better than chance, you can't even tell from its output.  Of course, Google Translate would not do this to you. But most Natural Language Processing libraries do, and it's terrible.
 
@@ -12,7 +12,7 @@ include ./meta.jade
 
     p Have a look through the #[a(href="http://gate.ac.uk/sale/tao/split.html") GATE software]. There's a lot there, developed over 12 years and many person-hours. But there's approximately zero curation.  The philosophy is just to provide things. It's up to you to decide what to use.
 
-    p This is bad. It's bad to provide an implementation of #[a(href="https://gate.ac.uk/sale/tao/splitch18.html") MiniPar], and have it just...sit there, with no hint that it's 20 years old and should not be used. The RASP parser, too. Why are these provided? Worse, why is there no warning? The #[a(href="http://webdocs.cs.ualberta.ca/~lindek/minipar.htm") Minipar homepage] puts the software in the right context:
+    p This is bad. It's bad to provide an implementation of #[a(href="https://gate.ac.uk/sale/tao/splitch18.html") MiniPar], and have it just...sit there, with no hint that it's 20 years old and should not be used. The RASP parser, too. Why are these provided? Worse, why is there no warning? The #[a(href="https://web.archive.org/web/20150907234221/http://webdocs.cs.ualberta.ca/~lindek/minipar.htm") Minipar homepage] puts the software in the right context:
     
     blockquote
         p MINIPAR is a broad-coverage parser for the English language. An evaluation with the SUSANNE corpus shows that MINIPAR achieves about 88% precision and 80% recall with respect to dependency relationships. MINIPAR is very efficient, #[strong on a Pentium II 300 with 128MB memory], it parses about 300 words per second.
@@ -23,13 +23,13 @@ include ./meta.jade
 
     h3 Why I didn't contribute to NLTK
 
-    p Various people have asked me why I decided to make a new Python NLP library, #[a(href="http://spacy.io") spaCy], instead of supporting the #[a(href="http://nltk.org") NLTK] project. This is the main reason. You can't contribute to a project if you believe that the first thing that they should do is throw almost all of it away. You should just make your own project, which is what I did.
+    p Various people have asked me why I decided to make a new Python NLP library, #[a(href="https://spacy.io") spaCy], instead of supporting the #[a(href="http://nltk.org") NLTK] project. This is the main reason. You can't contribute to a project if you believe that the first thing that they should do is throw almost all of it away. You should just make your own project, which is what I did.
     p Have a look through #[a(href="http://www.nltk.org/py-modindex.html") the module list of NLTK]. It looks like there's a lot there, but there's not. What NLTK has is a decent tokenizer, some passable stemmers, a good implementation of the Punkt sentence boundary detector (after #[a(href="http://joelnothman.com/") Joel Nothman] rewrote it), some visualization tools, and some wrappers for other libraries. Nothing else is of any use.
 
     p For instance, consider #[code nltk.parse]. You might think that amongst all this code there was something that could actually predict the syntactic structure of a sentence for you, but you would be wrong. There are wrappers for the BLLIP and Stanford parsers, and since March there's been an implementation of Nivre's 2003 transition-based dependency parser. Unfortunately no model is provided for it, as they rely on an external wrapper of an external learner, which is unsuitable for the structure of their problem. So the implementation is too slow to be actually useable.
 
-    p This problem is totally avoidable, if you just sit down and write good code, instead of stitching together external dependencies.  I pointed NLTK to my tutorial describing #[a(href="http://spacy.io/blog/parsing-english-in-python/") how to implement a modern dependency parser], which includes a BSD-licensed implementation in 500 lines of Python. I was told "thanks but no thanks", and #[a(href="https://github.com/nltk/nltk/issues/694") the issue was abruptly closed]. Another researcher's offer from 2012 to implement this type of model also went #[a(href="http://arxiv.org/pdf/1409.7386v1.pdf") unanswered].
+    p This problem is totally avoidable, if you just sit down and write good code, instead of stitching together external dependencies.  I pointed NLTK to my tutorial describing #[a(href="https://spacy.io/blog/parsing-english-in-python") how to implement a modern dependency parser], which includes a BSD-licensed implementation in 500 lines of Python. I was told "thanks but no thanks", and #[a(href="https://github.com/nltk/nltk/issues/694") the issue was abruptly closed]. Another researcher's offer from 2012 to implement this type of model also went #[a(href="http://arxiv.org/pdf/1409.7386v1.pdf") unanswered].
 
-    p The story in #[code nltk.tag] is similar. There are plenty of wrappers, for the external libraries that have actual taggers. The only actual tagger model they distribute is #[a(href="http://spacy.io/blog/part-of-speech-POS-tagger-in-python/") terrible].  Now it seems that #[a(href="https://github.com/nltk/nltk/issues/1063") NLTK does not even know how its POS tagger was trained]. The model is just this .pickle file that's been passed around for 5 years, its origins lost to time. It's not okay to offer this to people, to recommend they use it.
+    p The story in #[code nltk.tag] is similar. There are plenty of wrappers, for the external libraries that have actual taggers. The only actual tagger model they distribute is #[a(href="https://spacy.io/blog/part-of-speech-POS-tagger-in-python") terrible].  Now it seems that #[a(href="https://github.com/nltk/nltk/issues/1063") NLTK does not even know how its POS tagger was trained]. The model is just this .pickle file that's been passed around for 5 years, its origins lost to time. It's not okay to offer this to people, to recommend they use it.
 
     p I think open source software should be very careful to make its limitations clear. It's a disservice to provide something that's much less useful than you imply. It's like offering your friend a lift and then not showing up. It's totally fine to not do something &ndash; so long as you never suggested you were going to do it. There are ways to do worse than nothing. 
diff --git a/website/src/jade/blog/displacy/index.jade b/website/src/jade/blog/displacy/index.jade
index cac4fae5b..ded010a22 100644
--- a/website/src/jade/blog/displacy/index.jade
+++ b/website/src/jade/blog/displacy/index.jade
@@ -2,7 +2,7 @@ include ../../header.jade
 include ./meta.jade
 
 mixin Displacy(sentence, caption_text, height)
-    - var url = "http://api.spacy.io/displacy/?full=" + sentence.replace(" ", "%20")
+    - var url = "https://api.spacy.io/displacy/?full=" + sentence.replace(/\s+/g, "%20")
 
     .displacy
         iframe.displacy(src="/resources/displacy/robots.html" height=height)
@@ -20,7 +20,7 @@ mixin Displacy(sentence, caption_text, height)
 
     p A syntactic dependency parse is a kind of shallow meaning representation. It's an important piece of many language understanding and text processing technologies. Now that these representations can be computed quickly, and with increasingly high accuracy, they're being used in lots of applications &ndash; translation, sentiment analysis, and summarization are major application areas.
 
-    p I've been living and breathing similar representations for most of my career. But there's always been a problem: talking about these things is tough. Most people haven't thought much about grammatical structure, and the idea of them is inherently abstract.  When I left academia to write #[a(href="http://spaCy.io") spaCy], I knew I wanted a good visualizer. Unfortunately, I also knew I'd never be the one to write it. I'm deeply graphically challenged. Fortunately, when working with #[a(href="http://ines.io") Ines] to build this site, she really nailed the problem, with a solution I'd never have thought of. I really love the result, which we're calling #[a(href="http://api.spacy.io/displacy") displaCy]:
+    p I've been living and breathing similar representations for most of my career. But there's always been a problem: talking about these things is tough. Most people haven't thought much about grammatical structure, and the idea of them is inherently abstract.  When I left academia to write #[a(href="https://spacy.io") spaCy], I knew I wanted a good visualizer. Unfortunately, I also knew I'd never be the one to write it. I'm deeply graphically challenged. Fortunately, when working with #[a(href="http://ines.io") Ines] to build this site, she really nailed the problem, with a solution I'd never have thought of. I really love the result, which we're calling #[a(href="https://api.spacy.io/displacy") displaCy]:
     
         +Displacy("Robots in popular culture are there to remind us of the awesomeness of unbounded human agency", "Click the button to full-screen and interact, or scroll to see the full parse.", 325)
 
@@ -40,7 +40,7 @@ mixin Displacy(sentence, caption_text, height)
 
         p To me, this seemed like witchcraft, or a hack at best. But I was quickly won over: if all we do is declare the data and the relationships, in standards-compliant HTML and CSS, then we can simply step back and let the browser do its job. We know the code will be small, the layout will work on a variety of display, and we'll have a ready separation of style and content. For long output, we simply let the graphic overflow, and let users scroll.
 
-        p What I'm particularly excited about is the potential for displaCy as an #[a(href="http://api.spacy.io/displacy/?manual=Robots%20in%20popular%20culture%20are%20there%20to%20remind%20us%20of%20the%20awesomeness%20of%20unbounded%20human%20agency" target="_blank") annotation tool]. It may seem unintuitive at first, but I think it will be much better to annotate texts the way the parser operates, with a small set of actions and a stack, than by selecting arcs directly. Why? A few reasons:
+        p What I'm particularly excited about is the potential for displaCy as an #[a(href="https://api.spacy.io/displacy/?manual=Robots%20in%20popular%20culture%20are%20there%20to%20remind%20us%20of%20the%20awesomeness%20of%20unbounded%20human%20agency" target="_blank") annotation tool]. It may seem unintuitive at first, but I think it will be much better to annotate texts the way the parser operates, with a small set of actions and a stack, than by selecting arcs directly. Why? A few reasons:
 
         ul
             li You're always asked a question. You don't have to decide-what-to-decide.
diff --git a/website/src/jade/blog/displacy/meta.jade b/website/src/jade/blog/displacy/meta.jade
index e291fbc1e..ad0dd03e1 100644
--- a/website/src/jade/blog/displacy/meta.jade
+++ b/website/src/jade/blog/displacy/meta.jade
@@ -9,4 +9,4 @@
 - Meta.links[0].name = 'Reddit'
 - Meta.links[0].title = 'Discuss on Reddit'
 - Meta.links[0].url = "https://www.reddit.com/r/programming/comments/3hoj0b/displaying_linguistic_structure_with_css/"
-- Meta.image = "http://spacy.io/resources/img/displacy_screenshot.jpg"
+- Meta.image = "https://spacy.io/resources/img/displacy_screenshot.jpg"
diff --git a/website/src/jade/blog/eli5-computers-learn-reading/index.jade b/website/src/jade/blog/eli5-computers-learn-reading/index.jade
index d0a130fe1..d5b56ad7f 100644
--- a/website/src/jade/blog/eli5-computers-learn-reading/index.jade
+++ b/website/src/jade/blog/eli5-computers-learn-reading/index.jade
@@ -10,7 +10,7 @@ include ./meta.jade
 
     p It turns out that almost anything we say could mean many many different things, but we don't notice because almost all of those meanings would be weird or stupid or just not possible. If I say:
 
-    p.example #[a(href="http://api.spacy.io/displacy/?full=I%20saw%20a%20movie%20in%20a%20dress" target="_blank") I saw a movie in a dress]
+    p.example #[a(href="https://api.spacy.io/displacy/?full=I%20saw%20a%20movie%20in%20a%20dress" target="_blank") I saw a movie in a dress]
 
     p Would you ever ask me,
     
@@ -18,7 +18,7 @@ include ./meta.jade
     
     p It's weird to even think of that. But a computer just might, because there are other cases like:
 
-    p.example #[a(href="http://api.spacy.io/displacy/?full=The%20TV%20showed%20a%20girl%20in%20a%20dress" target="_blank") The TV showed a girl in a dress]
+    p.example #[a(href="https://api.spacy.io/displacy/?full=The%20TV%20showed%20a%20girl%20in%20a%20dress" target="_blank") The TV showed a girl in a dress]
 
     p Where the words hang together in the other way. People used to think that the answer was to tell the computer lots and lots of facts. But then you wake up one day and you're writing facts like #[em movies do not wear dresses], and you wonder where it all went wrong. Actually it's even worse than that. Not only are there too many facts, most of them are not even really facts! #[a(href="https://en.wikipedia.org/wiki/Cyc") People really tried this]. We've found that the world is made up of #[em if]s and #[em but]s.
 
diff --git a/website/src/jade/blog/eli5-computers-learn-reading/meta.jade b/website/src/jade/blog/eli5-computers-learn-reading/meta.jade
index 179c75ad6..5ebb973e2 100644
--- a/website/src/jade/blog/eli5-computers-learn-reading/meta.jade
+++ b/website/src/jade/blog/eli5-computers-learn-reading/meta.jade
@@ -3,7 +3,7 @@
 - Meta.headline = "Statistical NLP in Basic English"
 - Meta.description = "When I was little, my favorite TV shows all had talking computers. Now I’m big and there are still no talking computers, so I’m trying to make some myself. Well, we can make computers say things. But when we say things back, they don’t really understand. Why not?"
 - Meta.date = "2015-08-24"
-- Meta.url = "/blog/eli5-computers-learn-reading/"
+- Meta.url = "/blog/eli5-computers-learn-reading"
 - Meta.links = []
 //- Meta.links[0].id = 'reddit'
 //- Meta.links[0].name = "Reddit"
diff --git a/website/src/jade/blog/how-spacy-works/index.jade b/website/src/jade/blog/how-spacy-works/index.jade
index 0f6b05913..f3ec5c6f5 100644
--- a/website/src/jade/blog/how-spacy-works/index.jade
+++ b/website/src/jade/blog/how-spacy-works/index.jade
@@ -92,13 +92,13 @@ include ./meta.jade
 
     h3 Part-of-speech Tagger
 
-    p In 2013, I wrote a blog post describing #[a(href="/blog/part-of-speech-POS-tagger-in-python/") how to write a good part of speech tagger]. My recommendation then was to use greedy decoding with the averaged perceptron.  I think this is still the best approach, so it's what I implemented in spaCy.
+    p In 2013, I wrote a blog post describing #[a(href="/blog/part-of-speech-POS-tagger-in-python") how to write a good part of speech tagger]. My recommendation then was to use greedy decoding with the averaged perceptron.  I think this is still the best approach, so it's what I implemented in spaCy.
 
     p The tutorial also recommends the use of Brown cluster features, and case normalization features, as these make the model more robust and domain independent.  spaCy's tagger makes heavy use of these features.
 
     h3 Dependency Parser
 
-    p The parser uses the algorithm described in my #[a(href="/blog/parsing-english-in-python/") 2014 blog post].  This algorithm, shift-reduce dependency parsing, is becoming widely adopted due to its compelling speed/accuracy trade-off.
+    p The parser uses the algorithm described in my #[a(href="/blog/parsing-english-in-python") 2014 blog post].  This algorithm, shift-reduce dependency parsing, is becoming widely adopted due to its compelling speed/accuracy trade-off.
 
     p Some quick details about spaCy's take on this, for those who happen to know these models well.  I'll write up a better description shortly.
 
diff --git a/website/src/jade/blog/index.jade b/website/src/jade/blog/index.jade
index 702154ec7..a3acdc624 100644
--- a/website/src/jade/blog/index.jade
+++ b/website/src/jade/blog/index.jade
@@ -33,7 +33,7 @@ include ../header.jade
 
 +WritePage(Site, Authors.spacy, Page)
     section.intro.profile
-        p A lot of work has gone into #[strong spaCy], but no magic. We plan to keep no secrets. We want you to be able to #[a(href="/blog/spacy-now-mit") build your business] on #[strong spaCy] &ndash; so we want you to understand it. Tell us whether you do. #[span.social #[a(href="//twitter.com/" + Site.twitter, target="_blank") Twitter] #[a(href="mailto:contact@spacy.io") Contact us]]
+        p A lot of work has gone into #[strong spaCy], but no magic. We plan to keep no secrets. We want you to be able to #[a(href="/blog/spacy-now-mit") build your business] on #[strong spaCy] &ndash; so we want you to understand it. Tell us whether you do. #[span.social #[a(href="https://twitter.com/" + Site.twitter, target="_blank") Twitter] #[a(href="mailto:contact@spacy.io") Contact us]]
             nav(role='navigation')
                 ul
                     li #[a.button(href='#blogs') Blog]
diff --git a/website/src/jade/blog/introducing-spacy/index.jade b/website/src/jade/blog/introducing-spacy/index.jade
index 3d972e56c..263de4034 100644
--- a/website/src/jade/blog/introducing-spacy/index.jade
+++ b/website/src/jade/blog/introducing-spacy/index.jade
@@ -19,4 +19,4 @@ include ./meta.jade
             +TweetThis("Computers don't understand text. This is unfortunate, because that's what the web is mostly made of.", Meta.url)
 
         p If none of that made any sense to you, here's the gist of it.  Computers don't understand text.  This is unfortunate, because that's what the web almost entirely consists of.  We want to recommend people text based on other text they liked.  We want to shorten text to display it on a mobile screen.  We want to aggregate it, link it, filter it, categorise it, generate it and correct it.  
-        p spaCy provides a library of utility functions that help programmers build such products.  It's commercial open source software: you can either use it under the AGPL, or you can buy a commercial license under generous terms (Note: #[a(href="/blog/spacy-now-mit/") spaCy is now licensed under MIT]).
+        p spaCy provides a library of utility functions that help programmers build such products.  It's commercial open source software: you can either use it under the AGPL, or you can buy a commercial license under generous terms (Note: #[a(href="/blog/spacy-now-mit") spaCy is now licensed under MIT]).
diff --git a/website/src/jade/blog/spacy-now-mit/index.jade b/website/src/jade/blog/spacy-now-mit/index.jade
index 48db100a0..5723f9a78 100644
--- a/website/src/jade/blog/spacy-now-mit/index.jade
+++ b/website/src/jade/blog/spacy-now-mit/index.jade
@@ -5,7 +5,7 @@ include ../../header.jade
 +WritePost(Meta)
     //# AGPL not free enough: spaCy now under MIT, offering  adaptation as a service
 
-    p Three big announcements for #[a(href="http://spacy.io") spaCy], a Python library for industrial-strength natural language processing (NLP).
+    p Three big announcements for #[a(href="https://spacy.io") spaCy], a Python library for industrial-strength natural language processing (NLP).
     
     ol
         li The founding team is doubling in size: I'd like to welcome my new co-founder, #[a(href="https://www.linkedin.com/profile/view?id=ADEAAADkZcYBnipeHOAS6HqrDBPK1IzAAVI64ds&authType=NAME_SEARCH&authToken=YYZ1&locale=en_US&srchid=3310922891443387747239&srchindex=1&srchtotal=16&trk=vsrp_people_res_name&trkInfo=VSRPsearchId%3A3310922891443387747239%2CVSRPtargetId%3A14968262%2CVSRPcmpt%3Aprimary%2CVSRPnm%3Atrue%2CauthType%3ANAME_SEARCH") Henning Peters].
diff --git a/website/src/jade/docs/_spec.jade b/website/src/jade/docs/_spec.jade
index 08a31e810..263724633 100644
--- a/website/src/jade/docs/_spec.jade
+++ b/website/src/jade/docs/_spec.jade
@@ -52,7 +52,7 @@ details
 
   p The part-of-speech tagger uses the OntoNotes 5 version of the Penn Treebank tag set.  We also map the tags to the simpler Google Universal POS Tag set.
 
-  p Details #[a(href="https://github.com/honnibal/spaCy/blob/master/spacy/en/pos.pyx#L124") here].
+  p Details #[a(href="https://github.com/honnibal/spaCy/blob/master/spacy/tagger.pyx") here].
 
 details
   summary: h4 Lemmatization
diff --git a/website/src/jade/header.jade b/website/src/jade/header.jade
index f9da61a27..32deb5a6d 100644
--- a/website/src/jade/header.jade
+++ b/website/src/jade/header.jade
@@ -2,20 +2,20 @@
 - Site.name = "spaCy.io"
 - Site.slogan = "Build Tomorrow's Language Technologies"
 - Site.description = "spaCy is a library for industrial-strength text processing in Python. If you're a small company doing NLP, we want spaCy to seem like a minor miracle."
-- Site.image = "http://spacy.io/resources/img/social.png"
-- Site.image_small = "http://spacy.io/resources/img/social_small.png"
+- Site.image = "https://spacy.io/resources/img/social.png"
+- Site.image_small = "https://spacy.io/resources/img/social_small.png"
 - Site.twitter = "spacy_io"
-- Site.url = "http://spacy.io"
+- Site.url = "https://spacy.io"
 -
 - Authors = {"matt": {}, "spacy": {}};
 - Authors.matt.name = "Matthew Honnibal"
-- Authors.matt.bio = "Matthew Honnibal is the author of the <a href=\"http://spacy.io\">spaCy</a>  software and the sole founder of its parent company. He studied linguistics as an undergrad, and never thought he'd be a programmer. By 2009 he had a PhD in computer science, and in 2014 he left academia to found Syllogism Co. He's from Sydney and lives in Berlin."
+- Authors.matt.bio = "Matthew Honnibal is the author of the <a href=\"https://spacy.io\">spaCy</a>  software and the sole founder of its parent company. He studied linguistics as an undergrad, and never thought he'd be a programmer. By 2009 he had a PhD in computer science, and in 2014 he left academia to found Syllogism Co. He's from Sydney and lives in Berlin."
 
 - Authors.matt.image = "/resources/img/matt.png"
 - Authors.matt.twitter = "honnibal"
 - 
 - Authors.spacy.name = "SpaCy.io"
-- Authors.spacy.bio = "<a href=\"http://spacy.io\">spaCy</a> is a library for industrial-strength natural language processing in Python and Cython. It features state-of-the-art speed and accuracy, a concise API, and great documentation. If you're a small company doing NLP, we want spaCy to seem like a minor miracle."
+- Authors.spacy.bio = "<a href=\"https://spacy.io\">spaCy</a> is a library for industrial-strength natural language processing in Python and Cython. It features state-of-the-art speed and accuracy, a concise API, and great documentation. If you're a small company doing NLP, we want spaCy to seem like a minor miracle."
 - Authors.spacy.image = "/resources/img/social_small.png"
 - Authors.spacy.twitter = "spacy_io"
 
@@ -27,10 +27,11 @@
 -    Page.active[type] = true;
 -    Page.links = [];
 -    if (type == "home") {
--       Page.url = "";
+-       Page.url = "/";
 -    } else {
 -       Page.url = "/" + type;
 -    }
+-    Page.canonical_url = Site.url + Page.url;
 -
 -    // Set defaults
 -    Page.description = Site.description;
@@ -57,6 +58,7 @@
 -    Page.description = Meta.description
 -    Page.date = Meta.date
 -    Page.url = Meta.url
+-    Page.canonical_url = Site.url + Page.url;
 -    Page.active["blog"] = true
 -    Page.links = Meta.links
 -    if (Meta.image != null) {
@@ -98,6 +100,8 @@ mixin WritePage(Site, Author, Page)
             meta(property="og:site_name" content=Site.name)
             meta(property="article:published_time" content=getDate(Page.date).timestamp)
             link(rel="stylesheet" href="/resources/css/style.css")
+            if Page.canonical_url
+                link(rel="canonical" href=Page.canonical_url)
 
             //[if lt IE 9]><script src="//html5shiv.googlecode.com/svn/trunk/html5.js"></script><![endif]
 
@@ -114,7 +118,7 @@ mixin WritePage(Site, Author, Page)
             nav(role="navigation")
                 li(class={active: Page.active.home}): a(href="/") Home
                 li(class={active: Page.active.docs}): a(href="/docs") Docs
-                li: a(href="http://api.spacy.io/displacy", target="_blank") Demo
+                li: a(href="https://api.spacy.io/displacy", target="_blank") Demo
                 li(class={active: Page.active.blog}): a(href="/blog") Blog
             main#content
                 block
@@ -158,10 +162,10 @@ mixin WritePost(Meta)
                 +WriteAuthorBio(Author)
  
 mixin WriteByline(Author, Meta)
-    .subhead by #[a(href="//twitter.com/" + Author.twitter, rel="author" target="_blank") #{Author.name}] on #[time #{getDate(Meta.date).fulldate}]
+    .subhead by #[a(href="https://twitter.com/" + Author.twitter, rel="author" target="_blank") #{Author.name}] on #[time #{getDate(Meta.date).fulldate}]
 
 mixin WriteShareLinks(headline, url, twitter, links)
-    a.button.button-twitter(href="http://twitter.com/share?text=" + headline + "&url=" + Site.url + url + "&via=" + twitter title="Share on Twitter" target="_blank")
+    a.button.button-twitter(href="https://twitter.com/share?text=" + headline.replace(/\s+/g, "%20") + "&url=" + Site.url + url + "&via=" + twitter title="Share on Twitter" target="_blank")
         | Share on Twitter 
     if links
         .discuss
@@ -174,11 +178,11 @@ mixin WriteShareLinks(headline, url, twitter, links)
                         | Discuss on #{link.name}
 
 mixin TweetThis(text, url)
-    p #[span #{text} #[a.share(href='http://twitter.com/share?text="' + text + '"&url=' + Site.url + url + '&via=' + Site.twitter title='Share on Twitter' target='_blank') Tweet]]
+    p #[span #{text} #[a.share(href='https://twitter.com/share?text="' + text.replace(/\s+/g, "%20") + '"&url=' + Site.url + url + '&via=' + Site.twitter title='Share on Twitter' target='_blank') Tweet]]
 
 mixin WriteAuthorBio(Author)
     section.intro.profile
-        p #[img(src=Author.image)] !{Author.bio} #[span.social #[a(href="//twitter.com/" + Author.twitter target="_blank") Twitter]]
+        p #[img(src=Author.image alt=Author.name)] !{Author.bio} #[span.social #[a(href="https://twitter.com/" + Author.twitter target="_blank") Twitter]]
 
 
 - var getDate = function(input) {
diff --git a/website/src/jade/home/_online_demo.jade b/website/src/jade/home/_online_demo.jade
index 275dea2be..9feba0ed3 100644
--- a/website/src/jade/home/_online_demo.jade
+++ b/website/src/jade/home/_online_demo.jade
@@ -1,5 +1,5 @@
 mixin Displacy(sentence, caption_text, height)
-  - var url = "http://api.spacy.io/displacy/?full=" + sentence.replace(" ", "%20")
+  - var url = "https://api.spacy.io/displacy/?full=" + sentence.replace(/\s+/g, "%20")
 
   .displacy
     iframe.displacy(src="/resources/displacy/displacy_demo.html" height=height)
@@ -17,6 +17,6 @@ mixin Displacy(sentence, caption_text, height)
   275
 )
 
-p #[a(href="http://api.spacy.io/displacy") displaCy] lets you peek inside spaCy's syntactic parser, as it reads a sentence word-by-word. By repeatedly choosing from a small set of actions, it links the words together according to their syntactic structure. This type of representation powers a wide range of technologies, from translation and summarization, to sentiment analysis and algorithmic trading.  #[a(href="/blog/displacy") Read more.]
+p #[a(href="https://api.spacy.io/displacy") displaCy] lets you peek inside spaCy's syntactic parser, as it reads a sentence word-by-word. By repeatedly choosing from a small set of actions, it links the words together according to their syntactic structure. This type of representation powers a wide range of technologies, from translation and summarization, to sentiment analysis and algorithmic trading.  #[a(href="/blog/displacy") Read more.]
 
 
diff --git a/website/src/jade/home/_usage.jade b/website/src/jade/home/_usage.jade
index c87c01ad0..fac178f24 100644
--- a/website/src/jade/home/_usage.jade
+++ b/website/src/jade/home/_usage.jade
@@ -58,4 +58,4 @@ mixin example(name)
   ul
     li: a(href="/docs#api") API documentation
     li: a(href="/docs#tutorials") Tutorials
-    li: a(href="/docs/#spec") Annotation specs
+    li: a(href="/docs#spec") Annotation specs
diff --git a/website/src/jade/tutorials/add-a-language/meta.jade b/website/src/jade/tutorials/add-a-language/meta.jade
index 04818d968..b6fde298b 100644
--- a/website/src/jade/tutorials/add-a-language/meta.jade
+++ b/website/src/jade/tutorials/add-a-language/meta.jade
@@ -1,6 +1,6 @@
 - var Meta = {}
 - Meta.author_id = 'spacy'
-- Meta.headline = "Tutorial: Adding a language to spaCy"
+- Meta.headline = "Adding a language to spaCy"
 - Meta.description = "Long awaited documentation for adding a language to spaCy"
 - Meta.date = "2015-08-18"
 - Meta.url = "/tutorials/add-a-language"
diff --git a/website/src/jade/tutorials/index.jade b/website/src/jade/tutorials/index.jade
deleted file mode 100644
index e69de29bb..000000000
diff --git a/website/src/jade/tutorials/load-new-word-vectors/meta.jade b/website/src/jade/tutorials/load-new-word-vectors/meta.jade
new file mode 100644
index 000000000..ee5aa3a17
--- /dev/null
+++ b/website/src/jade/tutorials/load-new-word-vectors/meta.jade
@@ -0,0 +1,8 @@
+- var Meta = {}
+- Meta.headline = "Load new word vectors"
+- Meta.description = "Word vectors allow simple similarity queries, and drive many NLP applications. This tutorial explains how to load custom word vectors into spaCy, to make use of task or data-specific representations."
+- Meta.author_id = "matt"
+- Meta.date = "2015-09-24"
+- Meta.url = "/tutorials/load-new-word-vectors"
+- Meta.active = { "blog": true }
+- Meta.links = []
diff --git a/website/src/jade/tutorials/mark-adverbs/meta.jade b/website/src/jade/tutorials/mark-adverbs/meta.jade
index 259ac84f4..e6b4c4720 100644
--- a/website/src/jade/tutorials/mark-adverbs/meta.jade
+++ b/website/src/jade/tutorials/mark-adverbs/meta.jade
@@ -1,5 +1,5 @@
 - var Meta = {}
-- Meta.headline = "Tutorial: Mark all adverbs, particularly for verbs of speech"
+- Meta.headline = "Mark all adverbs, particularly for verbs of speech"
 - Meta.author_id = 'matt'
 - Meta.description = "Let's say you're developing a proofreading tool, or possibly an IDE for writers.  You're convinced by Stephen King's advice that adverbs are not your friend so you want to highlight all adverbs."
 - Meta.date = "2015-08-18"
diff --git a/website/src/jade/tutorials/syntax-search/meta.jade b/website/src/jade/tutorials/syntax-search/meta.jade
index cfdc755e9..126a671fd 100644
--- a/website/src/jade/tutorials/syntax-search/meta.jade
+++ b/website/src/jade/tutorials/syntax-search/meta.jade
@@ -1,5 +1,5 @@
 - var Meta = {}
-- Meta.headline = "Tutorial: Search Reddit for comments about Google doing something"
+- Meta.headline = "Search Reddit for comments about Google doing something"
 - Meta.description = "Example use of the spaCy NLP tools for data exploration. Here we will look for Reddit comments that describe Google doing something, i.e. discuss the company's actions. This is difficult, because other senses of \"Google\" now dominate usage of the word in conversation, particularly references to using Google products."
 - Meta.author_id = "matt"
 - Meta.date = "2015-08-18"
diff --git a/website/src/jade/tutorials/twitter-filter/index.jade b/website/src/jade/tutorials/twitter-filter/index.jade
index 2a6f19823..c87b2d8f1 100644
--- a/website/src/jade/tutorials/twitter-filter/index.jade
+++ b/website/src/jade/tutorials/twitter-filter/index.jade
@@ -4,7 +4,7 @@ include ./meta.jade
 
 +WritePost(Meta)
     section.intro
-        p #[a(href="http://spaCy.io") spaCy] is great for data exploration. Poking, prodding and sifting is fundamental to good data science. In this tutorial, we'll do a broad keword search of Twitter, and then sift through the live stream of tweets, zooming in on some topics and excluding others.
+        p #[a(href="https://spacy.io") spaCy] is great for data exploration. Poking, prodding and sifting is fundamental to good data science. In this tutorial, we'll do a broad keword search of Twitter, and then sift through the live stream of tweets, zooming in on some topics and excluding others.
 
         p An example filter-function:
 
diff --git a/website/src/jade/tutorials/twitter-filter/meta.jade b/website/src/jade/tutorials/twitter-filter/meta.jade
index 2d989d595..c705e597b 100644
--- a/website/src/jade/tutorials/twitter-filter/meta.jade
+++ b/website/src/jade/tutorials/twitter-filter/meta.jade
@@ -1,5 +1,5 @@
 - var Meta = {}
-- Meta.headline = "Tutorial: Finding Relevant Tweets"
+- Meta.headline = "Finding Relevant Tweets"
 - Meta.author_id = 'matt'
 - Meta.description = "In this tutorial, we will use word vectors to search for tweets about Jeb Bush. We'll do this by building up two word lists: one that represents the type of meanings in the Jeb Bush tweets, and another to help screen out irrelevant tweets that mention the common, ambiguous word 'bush'."
 - Meta.date = "2015-08-18"