hippmem-py

Python bindings for HIPPMEM — native associative memory engine for AI agents.

Ecosystem: Rust engine · Python bindings · MCP server

Install

pip install hippmem

No GPU, API key, or network required — the deterministic fallback backend works offline.

Quick Start

from hippmem import Engine

# Hash embedder — offline, zero config (default)
engine = Engine.open()
engine.write("The user prefers Rust.", content_type="Preference")
engine.write("The user chose redb — pure Rust, fast compile.", content_type="Decision")

results = engine.retrieve("Why did the user choose redb?", top_k=3)
for r in results.results:
    print(f"[{r.score:.3f}] {r.content}")
    print(f"  dimensions: {r.dimensions}")

engine.close()

Neural embedder (higher semantic accuracy)

engine = Engine.open(
    embedder="neural",
    api_base_url="https://api.openai.com/v1",
    api_key="sk-...",
    model="text-embedding-3-small",
)

Why associative memory?

Most memory solutions are vector databases — store embeddings, search by similarity. HIPPMEM discovers associations between memories at write time (entity, causal, temporal, semantic, topic) and retrieves via spreading activation over typed edges. The result: it remembers WHY, not just WHAT.

Documentation

For deeper architecture and algorithm details, see the main HIPPMEM documentation.

Development

git clone https://github.com/hippmem/hippmem-py.git
cd hippmem-py
python3 -m venv .venv && source .venv/bin/activate
pip install maturin pytest
maturin develop
pytest

License

Apache 2.0. See LICENSE and COPYRIGHT.

The underlying hippmem-engine is AGPL-3.0-only. Importing hippmem-py does NOT subject your program to AGPL — only the engine crate itself is AGPL. See the HIPPMEM licensing overview.