Embedded Graph Database for Python
Graphs · Vectors · Search
pip install cogdb
No server. No config. Runs in notebooks, apps, even your browser.
from cog.torque import Graph
g = Graph("cities")
g.put("Tokyo", "is_a", "city")
g.put("Tokyo", "known_for", "sushi")
g.put("Tokyo", "country", "Japan")
g.put("Naples", "is_a", "city")
g.put("Naples", "known_for", "espresso")
g.put("Naples", "country", "Italy")
# auto-embed all nodes for semantic search
g.vectorize()
# Where should I go for coffee?
g.v().k_nearest("coffee", k=1).inc("known_for").all()Result
['Naples']Zero Setup
pip install cogdb — no server, no Docker, no config, no dependencies.
Runs Anywhere
Jupyter notebooks, Python apps, Google Colab, even the browser via Pyodide.
Built-in Storage Engine
No third-party dependencies. Your data survives restarts.
Python-Native Query API
Traverse graphs with Torque — a fluent, chainable Python API. No Cypher, no query strings.
Graph + Vector Search
Combine traversal with similarity — filter by relationships and meaning.
Visualization
Built-in graph visualization. See your data as you build.
Stay in the Loop
Get notified about new features and releases