Skillget

magnet-gateway

agent
Security Audit
Fail
Health Warn
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 6 GitHub stars
Code Fail
  • eval() — Dynamic code execution via eval() in sdk/magnet/buffer.py
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

Give your AI a memory. Drop-in learning layer for any LLM — no retraining, no RAG setup.

README.md

Magnet

Documentation License: MIT Built by Agent Magnet PyPI Stars Last Commit

Your AI forgets every user the moment the session ends.
Magnet fixes that — without changing your code.

How It Works

User sends message → Magnet injects memory → LLM responds → Magnet learns

  • Learns from corrections, not just conversations
  • Builds a profile that gets smarter with every interaction
  • Compresses thousands of messages into a lightweight JSON snapshot

Quick Start

Proxy Mode (2 steps)

# Step 1: Start services
docker compose up -d

# Step 2: Change your base URL — nothing else
import openai

client = openai.Client(
    base_url="http://localhost:8000/v1",
    api_key="your-openai-api-key"
)

response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello!"}],
    extra_headers={"x-session-id": "user-123"}  # Magnet tracks memory per user
)

SDK Mode

pip install git+https://github.com/helinakdogan/magnet-gateway#subdirectory=sdk

from magnet import BehavioralMemory

memory = BehavioralMemory(reflector_model="openai/gpt-4o-mini")

# Get context for a user
context = memory.get_injection(user_id="alice")

# Add a conversation to memory
memory.add(messages, user_id="alice")

Why Magnet

Traditional RAG Magnet
Setup Vector DB + embeddings + retrieval pipeline ✅ Drop-in proxy or one import
Latency Adds retrieval roundtrip on every call ✅ O(1) injection, async learning
Learning Static — you update it manually ✅ Adapts from every interaction
Privacy Shared embedding pool ✅ Per-user, self-hosted, no data sharing

Architecture

Your AI remembers what matters across three layers — each one builds on the last.

Layer 1 — Redis (always on, real-time preferences and corrections)
Layer 2 — Qdrant (episodic recall, semantic memory from past sessions)
Layer 3 — Neo4j (relationships and long-term knowledge graph)

Configuration

Set these in your .env file:

Variable Description
OPENAI_API_KEY Used by the reflector model to analyze interactions.
REDIS_URL e.g. redis://localhost:6379. Used for Layer 1.
QDRANT_URL Used for Layer 2 episodic memory.
NEO4J_URL Used for Layer 3 graph knowledge.

Documentation

Full docs at agentmagnet.app/docs:

Section What's Covered
Quickstart Install → setup → first interaction in 2 minutes
Architecture Details on the 3-layer memory engine
Proxy Mode How to use Magnet as a transparent gateway
SDK Usage Deep integration into Python applications
Self Hosting Instructions for running Redis, Qdrant, and Neo4j

Contributing

Open an issue or submit a pull request — check CONTRIBUTING.md for guidelines.

If Magnet saved you from a bad context window, give it a ⭐

License

MIT — see LICENSE. Built by Agent Magnet.

Reviews (0)

No results found