Long-term memory for AI assistants. Graph + vector. Runs on your hardware.
Your AI forgets between sessions. RAG dumps documents that look similar. Vector databases match keywords but miss meaning. None of them learn.
AutoMem stores typed relationships and embeddings. When you ask "why did we choose PostgreSQL?", recall returns not just the matching memory — but the alternatives you considered, the principle behind the choice, and the related decisions that came after.
It scores 89.27% on the LoCoMo long-term memory benchmark (ACL 2024) judge-off, and 87.56% judge-on. See benchmarks/EXPERIMENT_LOG.md for methodology and history.
Additional LongMemEval and BEAM validation is tracked in benchmarks/EXPERIMENT_LOG.md; BEAM is currently reported as exploratory because published comparisons are not yet apples-to-apples.
| Use AutoMem if... | Look elsewhere if... |
|---|---|
| You want one memory across Claude / Cursor / ChatGPT / Codex | You need SOC2 / HIPAA audit logs and row-level ACLs |
| You're comfortable self-hosting (Docker or Railway) | You want a managed SaaS with a polished dashboard |
| You're a solo dev, prosumer, or small team | You're running a multi-agent swarm needing per-agent memory isolation |
| You want to own your memory data | You need an enterprise SLA and dedicated support |
If your row is on the right, AutoMem isn't it — yet. Try Mem0, Letta, or Zep instead.
AutoMem combines two storage layers behind a single API:
- FalkorDB stores memories as nodes with 11 typed relationships between them. The graph is the canonical record.
- Qdrant stores an embedding for every memory. Recall is a hybrid query — semantic similarity, graph traversal, temporal alignment, tag overlap, and importance — ranked by a 9-component score.
flowchart TB
subgraph service [AutoMem Service Flask]
API[REST API<br/>Memory Lifecycle]
Enrichment[Background Enrichment<br/>Pipeline]
Consolidation[Consolidation<br/>Engine]
Backups[Automated Backups<br/>Optional]
end
subgraph storage [Dual Storage Layer]
FalkorDB[(FalkorDB<br/>Graph Database)]
Qdrant[(Qdrant<br/>Vector Database)]
end
Client[AI Client] -->|Store/Recall/Associate| API
API --> FalkorDB
API --> Qdrant
Enrichment -->|11 edge types<br/>Pattern nodes| FalkorDB
Enrichment -->|Semantic search<br/>1024-d vectors| Qdrant
Consolidation --> FalkorDB
Consolidation --> Qdrant
Backups -.->|Optional| FalkorDB
Backups -.->|Optional| Qdrant
If Qdrant is unavailable, the graph still serves recall in a degraded mode. If FalkorDB is down, the API returns 503 — the graph is the source of truth.
Ask "why boring tech for Kafka?" and AutoMem doesn't just match the word "Kafka". It traverses the graph from the seed memories to find the bridge that connects them:
- Seed 1: "Migrated to PostgreSQL for operational simplicity"
- Seed 2: "Evaluating Kafka vs RabbitMQ for message queue"
- Bridge: "Team prefers boring technology — proven, debuggable systems"
Both seeds carry an EXEMPLIFIES edge to the bridge memory. AutoMem ranks the bridge above the seeds and surfaces it in the recall response, so the assistant answers with your reasoning, not isolated facts. Tune via expand_relations, relation_limit, and expansion_limit on GET /recall.
| Type | Use case | Example |
|---|---|---|
RELATES_TO |
General connection | Bug report → Related issue |
LEADS_TO |
Causal relationship | Problem → Solution |
OCCURRED_BEFORE |
Temporal sequence | Planning → Execution |
PREFERS_OVER |
User preferences | PostgreSQL → MongoDB |
EXEMPLIFIES |
Pattern examples | Code review → Best practice |
CONTRADICTS |
Conflicting info | Old approach → New approach |
REINFORCES |
Supporting evidence | Decision → Validation |
INVALIDATED_BY |
Outdated info | Legacy docs → Current docs |
EVOLVED_INTO |
Knowledge evolution | Initial design → Final design |
DERIVED_FROM |
Source tracking | Implementation → Spec |
PART_OF |
Hierarchical structure | Feature → Epic |
Three more edge types are added automatically by the enrichment pipeline and consolidation engine: SIMILAR_TO, PRECEDED_BY, and DISCOVERED.
AutoMem implements biological memory consolidation cycles. Wrong rabbit holes fade naturally. Important memories with strong connections strengthen over time.
| Cycle | Frequency | Purpose |
|---|---|---|
| Decay | Daily | Exponential relevance scoring (age, access, connections, importance) |
| Creative | Weekly | REM-like processing that discovers non-obvious connections |
| Cluster | Monthly | Groups similar memories, generates meta-patterns |
| Forget | Off by default | Archives low-relevance memories (<0.2), deletes very old (<0.05) |
Tune intervals via CONSOLIDATION_*_INTERVAL_SECONDS. See docs/ENVIRONMENT_VARIABLES.md.
For more on the recall scoring formula, enrichment internals, and how AutoMem differs from RAG and pure vector databases, see docs/COMPARISON.md.
AutoMem implements techniques from peer-reviewed memory research:
- HippoRAG 2 (Ohio State, 2025) — graph + vector hybrid for associative memory
- A-MEM (2025) — Zettelkasten-inspired dynamic memory organization
- MELODI (DeepMind, 2024) — gist-based memory compression
- ReadAgent (DeepMind, 2024) — episodic memory for context extension
Full writeups, findings, and how AutoMem implements each → docs/RESEARCH.md.
Deploys four services from pre-built Docker images: automem (the API), falkordb (graph), qdrant (vectors), and mcp-automem (the MCP bridge for ChatGPT, Claude.ai, and ElevenLabs). Auto-redeploys nightly on :stable. Roughly $0.50/month after the $5 free trial.
→ Full setup: INSTALLATION.md
git clone https://github.com/verygoodplugins/automem.git
cd automem
make dev| Service | URL | Purpose |
|---|---|---|
| AutoMem API | http://localhost:8001 |
Memory REST API |
| FalkorDB | localhost:6379 |
Graph database |
| Qdrant | localhost:6333 |
Vector database |
| FalkorDB Browser | http://localhost:3000 |
Local graph inspection UI |
→ Full setup: INSTALLATION.md
make install
source .venv/bin/activate
PORT=8001 python app.pyRequires Python 3.10+ (3.12 recommended). → INSTALLATION.md
| Client | Mode | Setup |
|---|---|---|
| Claude Desktop, Cursor, Claude Code, Codex, Copilot, Antigravity | Local MCP bridge | npx @verygoodplugins/mcp-automem setup |
| ChatGPT (developer mode), Claude.ai web/mobile, ElevenLabs Agents | Remote MCP (HTTPS) | docs/MCP_SSE.md |
| Anything else | Direct REST API | docs/API.md |
The MCP bridge is published as @verygoodplugins/mcp-automem. It handles client-specific config (rules files, hooks, templates) and proxies to your AutoMem service — local or Railway.
Direct API call:
import requests
token = "your-automem-api-token"
requests.post(
"https://your-automem.railway.app/memory",
headers={"Authorization": f"Bearer {token}"},
json={
"content": "Chose PostgreSQL over MongoDB for ACID compliance",
"type": "Decision",
"tags": ["database", "architecture"],
"importance": 0.9,
},
)Screenshots will be added once the referenced in-repo image assets are available.
Setup
- Installation guide — Railway, Docker, development
- Qdrant setup — vector database configuration
- Environment variables — full reference
API and integration
- API reference — endpoints, scoring, enrichment
- Remote MCP — ChatGPT, Claude.ai, ElevenLabs
- Migrations — moving from MCP SQLite
Research and comparison
- Research foundation — papers and how AutoMem implements them
- Comparison — vs. RAG, vector DBs, building your own
- Benchmark history — LoCoMo, LongMemEval, and BEAM methodology and runs
Operations
- Health monitoring & backups
- Testing guide — unit, integration, benchmarks
Community
- automem.ai — official site
- Discord — community chat
- X / @automem_ai — updates
- YouTube / @AutoJackBot — tutorials
- GitHub issues — bugs and feature requests
Sibling repos
mcp-automem— universal MCP bridge / install funnelautomem-evals— exploratory recall-quality labautomem-graph-viewer— standalone graph visualization
MIT licensed. Deploy anywhere. No vendor lock-in.