Which memory layer should I use?

Answer a few questions and get three picks from 191 catalogued memory systems. Every reason is a catalogued fact with its source. Picks are ranked on fit; GitHub stars only break ties. Archived and shut-down systems are left out. Where the catalog can't confirm a requirement, the system goes under might fit instead of being recommended.

What are you building?
Where should it run?
When a fact changes ("I moved to Hong Kong"), what should memory do?
What will it mostly remember?
Must-haves

Top picks

1 Mem0 Dedicated memory layers

Universal memory layer for AI agents. Three concurrent stores (vector + graph + KV); LLM-extracted facts; concurrent retrieval via ThreadPoolExecutor.

  • Licence: Apache-2.0 ↗
  • Deployment: Both ↗
  • Pricing: Free + paid ↗
  • Recently released: v2.2.1 (2026-09-25) — mem0ai on PyPI ↗
  • Adoption (tiebreaker): mem0ai/mem0 — ~67.3k stars, last commit 2026-09+ ↗
2 Hindsight (Vectorize) Dedicated memory layers

Standalone memory service from Vectorize. Open source. Biomimetic four-network design (World, Bank, Observation, Opinion). Ships an MCP memory server.

  • Licence: MIT ↗
  • Deployment: Both — managed Hindsight Cloud; Enterprise single-tenant (Vectorize cloud, customer AWS/GCP/Azure, or on-prem); self-host via Docker, Helm or pip (Postgres + pgvector) ↗
  • Pricing: OSS self-host free (MIT); Hindsight Cloud pay-as-you-go credits ($5 free credit on GitHub sign-up); Enterprise single-tenant custom ↗
  • Recently released: v1.1.0 (2026-10-21) ↗
  • Adoption (tiebreaker): vectorize-io/hindsight — ~44.6k stars, last commit 2026-10 ↗
3 Zep & Graphiti Dedicated memory layers

Bi-temporal knowledge graph (event time + ingestion time). Strong on chronological reasoning and contradiction tracking. Graphiti is the open-source core.

  • Licence: Apache-2.0 ↗
  • Deployment: Both — Zep Cloud managed; Graphiti OSS self-hosted (Neo4j/FalkorDB) ↗
  • Pricing: Free 10k credits/mo; Flex $125/mo ($104 annual) 50k credits; Flex Plus $375/mo ($312 annual) 200k credits; Enterprise custom ↗
  • Recently released: v0.30.0 (2026-09-01) (graphiti-core; MCP server mcp-v1.1.0 same day) ↗
  • Adoption (tiebreaker): getzep/graphiti — ~32k stars, last commit 2026-09 ↗

Also consider

137 systems considered · 8 ruled out by your must-haves or because they're archived / shut down.

Compare side by side

Compare · Self-hostManagedOpen sourceMCPConflict handlingMemory modelLicence
Mem0✓✓✓✓llm-arbitratevector, graph, kv67k2026-09-25Apache-2.0
Hindsight (Vectorize)✓✓✓✓nonevector, episode, summary45k2026-10-21MIT
Zep & Graphiti✓✓✓✓bi-temporalgraph, episode, fact32k2026-09-01Apache-2.0
Supermemory✓✓✓✓llm-arbitrategraph, vector, document31k2026-10-12MIT
Cognee✓✓✓✓llm-arbitrategraph, fact, document31k2026-10-16Apache-2.0
Letta / MemGPT✓✓✓✓llm-arbitratekv, vector, episode3.7k2026-10-29Apache-2.0
SuperLocalMemory✓✓✓✓nonevector, graph, fact2272026-10-17AGPL-3.0-or-later (commercial…
n8n AI Agent Memory✓✓—✓append-onlyvector, relational, episode208k2026-10-19Sustainable Use License + n8n…
Memori✓✓✓✓appendrelational, vector, fact17k2026-05-28Apache-2.0
Memvid✓✓✓✓appendfile, chunk, document17k2026-05-27Apache-2.0
MemMachine✓✓✓✓nonevector, graph, episode3.2k2026-05-18Apache-2.0
OMEGA✓—✓✓nonehybrid, fact2242026-10-08Apache-2.0
VEKTOR Memory✓✓—✓llm-arbitraterelational, vector, fact52026-07-16proprietary (commercial licence)
MemoraX AI✓?✓?noneparametric, weight2026-10-21MIT (MemoraX Code client)…
Aleph Alpha PhariaAI / PhariaCatch✓✓??llm-arbitratevector, file, document
Interloom✓✓??noneproprietary, skill
NAVER HyperCLOVA X THINK—✓??parametric, kv-cache, kv-token
LlamaIndex Memory✓✓✓✓append-onlykv, episode, summary53k2026-09-21MIT
CrewAI Memory✓✓✓✓llm-arbitratevector, kv, episode51k2026-10-14MIT
LangGraph Persistence✓✓✓✓last-write-winsrelational, kv, episode43k2026-10-16MIT
Agno (Phidata) Memory✓✓✓✓overwriterelational, vector, episode43k2026-10-21Apache-2.0
smolagents Memory✓✓✓✓appendkv, episode29k2026-05-29Apache-2.0
OpenAI Agents SDK Memory✓✓✓✓overwritekv, vector, episode29k2026-10-15MIT
Semantic Kernel Memory✓✓✓✓overwritevector, document, chunk28k2026-10-06MIT
Mastra Memory✓✓✓✓nonerelational, vector, episode27k2026-10-21Apache-2.0 (core; ee/ dirs…
Google ADK Memory✓✓✓✓overwritekv, vector, episode21k2026-08-07Apache-2.0
Memobase✓✓✓✓nonerelational, vector, profile2.8k2025-11-08Apache-2.0
Memoria (MatrixOrigin)✓✓✓✓nonehybrid, fact607Apache-2.0
NeoCognition????nonehybrid, episode, fact
OpenViking (ByteDance Volcengine)✓?✓✓file, vector, summary41k2026-10-22AGPL-3.0 (CLI/examples…
gbrain (Garry Tan)✓?✓✓human-arbitratefile, relational, vector31k2026-10-22MIT
TencentDB Agent Memory✓?✓?document-versionvector, graph, fact30k2026-10-21MIT
AutoGLM (Zhipu AI)✓✓✓?file, vector, episode26kApache-2.0
Botpress LLMz✓✓✓✓llm-arbitratevector, document, episode15kMIT (SDK/integration packages…
memU (NevaMind)✓?✓?agentfile, relational, vector15k2026-10-16Apache-2.0
MemoryBear✓?✓?human-arbitrategraph, relational, vector8.3k2026-10-14Apache-2.0
Honcho (Plastic Labs)✓?✓✓llm-arbitraterelational, vector, profile7.7k2026-10-21AGPL-3.0
Memanto (Moorcheh)✓?✓?human-arbitrate, bi-temporalproprietary, vector, fact2.3k2026-10-21MIT
M3-Agent (ByteDance)✓✓✓?graph, vector, episode1.4kApache-2.0
Neo4j Agent Memory (Neo4j Labs)✓?✓✓nonegraph, vector, entity5922026-10-22Apache-2.0
O-Mem✓✓✓?llm-arbitratevector, graph, fact79Apache-2.0
Backboard✓✓—?nonevector, episode18Unlicensed
Pieces for Developers (LTM)✓✓?✓vector, episode113
AI Singapore SEA-LION✓✓?—vector, chunk
Anthropic Claude Memory—✓?✓manualfile, fact, summary
Google Gemini Memory—✓?✓llm-arbitratekv, fact, profile
Grok Memory (xAI)—✓??llm-arbitratekv, fact
Hyperspell✓✓?✓nonevector, document, fact
LinkedIn Cognitive Memory Agent—✓??nonevector, graph, episode
Meta AI Memory—✓??llm-arbitratekv, fact
Microsoft Copilot Memory—✓?✓overwriterelational, fact
Mistral Agents API✓✓?✓nonekv, episode
Nyne✓✓??nonegraph, profile, fact
OpenAI ChatGPT Memory—✓?✓llm-arbitratekv, fact
Perplexity Memory—✓?✓llm-arbitratekv, fact
Relevance AI Memory—✓?✓overwritekv, fact
Trace✓✓??nonegraph, fact, episode
Voiceflow Memory—✓?✓last-write-winskv, fact
Langflow Memoryacquired✓✓✓✓append-onlyrelational, vector, episode156k2026-08-19MIT
Haystack Memory (deepset)✓✓✓✓appendkv, document26k2026-10-21Apache-2.0
Pydantic-AI + Hindsight✓✓✓✓nonevector, episode20k2026-04-15MIT
Memori (GibsonAI) (now Memori Labs)✓✓✓✓kv, fact17k2026-05-28Apache-2.0
Spring AI ChatMemory✓✓✓✓appendkv, document9.4k2026-08-21Apache-2.0
OpenMemory (now LongMemory)✓✓✓✓nonevector, graph, fact4.5k2025-12-12Apache-2.0
AGiXT Adaptive Memory✓✓✓?llm-arbitratevector, episode, fact3.2k2026-04-08MIT
LangMem (LangChain)✓✓✓✓llm-arbitratevector, kv, episode1.6k2025-10-27MIT
MemWal (now Walrus Memory)✓—✓?appendproprietary, document, episode117Apache-2.0
Cellon (Memory.inc)????nonevector, fact, episode
Hermes Agent memory (Nous Research)✓?✓✓agentfile, relational, fact256k2026-10-22MIT
LightRAG✓?✓?graph, vector, fact39k2026-09-02MIT
LightRAG✓?✓?graph, chunk39k2026-09-02MIT
GraphRAG (Microsoft)maintenance✓?✓?graph, vector, document35k2026-10-08MIT
LazyGraphRAG✓?✓?graph, chunk35k2026-08-21MIT
Praison Memory✓?✓✓vector, episode9.0k2026-10-19MIT
Strands Agents Memory (AWS)✓✓✓✓llm-arbitratevector, graph, fact8.9kApache-2.0
m_flow (M-Flow)✓?✓✓hybrid, graph, vector4.5k2026-08-03Apache-2.0
memory-lancedb-pro (OpenClaw plugin)✓?✓?llm-arbitratecolumn, vector, file4.5k2026-10-21MIT
NVIDIA Inference Context Memory Storage Platform (now NVIDIA CMX Context Memory Storage Platform)✓✓—?kv, tokenproprietary (hardware/DOCA) +…
AWS Bedrock AgentCore Memory—✓?✓noneproprietary, fact
Cloudflare Agent Memory—✓?✓nonehybrid, episode
Memory Store (now MemStore)✓✓?✓nonevector, episode, fact
Mistral Le Chat — Memories (now Vibe)✓✓?✓llm-arbitrategraph, kv, fact
Retell AI✓✓?✓nonekv, vector, episode
Vapi (voice agents)✓✓?✓overwritekv, fact
Memaryinactive✓✓✓?nonegraph, vector, episode2.6k2024-10-22MIT
DSPy History✓✓✓?kv, document37kMIT
Pydantic-AI MemoryTool✓✓✓✓noneproprietary, fact20kMIT
RAFT✓?✓?parametric, document13kApache-2.0
BGE-M3✓?✓?vector, chunk12k2026-08-24MIT
KAG✓?✓?graph, chunk9.0k2025-06-28Apache-2.0
AutoRAG✓?✓?hybrid, chunk5.1k2026-10-04Apache-2.0
Rivet Memory Node (Ironclad)✓✓✓?overwritekv, kv-token4.7kMIT
HippoRAG / HippoRAG2✓?✓✓graph, chunk4.0k2025-06-24MIT
HippoRAG 2✓?✓?graph, vector, fact4.0kMIT
CAG (Cache-Augmented Generation)✓?✓?kv-cache, document1.6kMIT
Inngest AgentKit — Network State✓✓✓✓overwritekv, kv-token926Apache-2.0
Memoripy✓?✓✓vector, fact, episode6952024-11-28Apache-2.0
HiRAG✓?✓?graph, chunk557MIT
Generative Semantic Workspace (GSW)????llm-arbitratevector, episode25
Anthropic Managed Agents — Memory—✓?✓nonefile
Microsoft Foundry Agent Service Memory??—✓llm-arbitrateproprietary, vector, profileproprietary
Perplexity Brain??—?proprietary, factproprietary
Vertex AI Memory Bank (now Agent Platform Memory Bank)—✓?✓nonevector, fact
AutoGen Memorymaintenance✓✓✓✓appendkv, vector, episode60k2025-09-30MIT
MemoRAG✓?✓?vector, fact2.3kApache-2.0
RAPTOR✓?✓?vector, chunk1.8kMIT
RETRO??—?vector, chunk8782023-10-30not applicable — no official…
FLARE✓?✓?vector, chunk668MIT
Atlas??—?vector, document560CC-BY-NC 4.0
PathRAG✓?✓?graph, fact376MIT
StructRAG??—?hybrid, chunk171no-license
TC-RAG??—?hybrid, chunk73no-license
ITER-RETGEN??—?vector, chunknot applicable — no official…
RAPTOR✓?✓?vector, chunk1.8kMIT
REALM??✓?vector, document1.8kApache-2.0
HyDE??—?vector, document582Unlicensed
Adaptive-RAG✓?✓?vector, chunk412Apache-2.0
PathRAG✓?✓?graph, chunk376MIT
Multi-Head RAG (MHRAG)????vector, chunk242
GraphRAFT????graph, document7
Stack AIacquired✓✓?✓vector, document
Beyond RAG for Agent Memory????llm-arbitratehybrid, chunk
ChunkRAG????vector, chunk
RouteRAG????hybrid, chunk
Speculative RAG????vector, chunk
WeKnow-RAG????hybrid, chunk
Self-RAG??✓?vector, chunk2.4kMIT
CRAG??—?vector, chunk469no-license
LongRAG??✓?vector, document248MIT

✓ confirmed · — no · ? not in the catalog yet. Hover a cell for the catalogued text. Already using some of these? Check your stack for archived or acquired dependencies.