Skip to main content

Remembr

Persistent memory infrastructure for AI agents.
Store, search, and delete memories across sessions — with one command to self-host.

Get Started →View on GitHub

pip install remembr·npm install @remembr/sdk

Quick Start

Add persistent memory to your agent in under 5 minutes.

8 Framework Adapters

Native adapters for LangChain, LangGraph, CrewAI, AutoGen, LlamaIndex, Pydantic AI, OpenAI Agents, and Haystack.

API Reference

Every endpoint, every parameter, every response shape documented.

Self-Hosting

One command. No API keys required. Runs fully local with Ollama embeddings.

Hybrid Search

Semantic + BM25 + recency scoring with configurable weights.

GDPR Compliant

Soft deletes, hard deletes, export, and targeted erasure at every scope level.

Dead simple API

from remembr import RemembrClient

async with RemembrClient(api_key="...", base_url="http://localhost:8000/api/v1") as client:
    session = await client.create_session(metadata={"user": "alice"})

    await client.store(
        "Alice prefers weekly billing summaries on Fridays.",
        role="user",
        session_id=session.session_id,
        tags=["kind:preference", "topic:billing"],
    )

    results = await client.search(
        "When should billing summaries be sent?",
        session_id=session.session_id,
        search_mode="hybrid",
    )
    print(results.results[0].content)