embedded graph datastore
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.




Graphlite is an MIT-licensed graph DB for Python.

There are currently no embedded graph databases for Python. Graphlite aims to change that by building a simple and fast graph layer over SQLite. Similar to FlockDB, Graphlite only stores adjacency lists, but they can be queried in the style of normal graph databases, e.g. with traversals.

>>> import graphlite as g
>>> db = g.connect(':memory:', graphs=['knows'])

>>> with db.transaction() as t:
...     for person in [2, 3]:
...         t.store(g.V(1).knows(person))

>>> db.find(g.V(1).knows).to(list)
[2, 3]

Graphlite inherits it's API from that of FlockDB's. Also, like FlockDB Graphlite only stores the relations between the nodes, which are integers- not the data of the nodes themselves. You will therefore need a high performance database like BerkleyDB or one of the dbm implementations to store your data.


  • Small, easy to learn API with around 100% coverage
  • Lazy generator based API when querying
  • Highly documented codebase


To install Graphlite, simply:

$ pip install graphlite