#1919th of 20 in Memory
memori
Long-term memory layer for LLM agents, with Python and TypeScript SDKs
- Stars
- 17.6k
- License
- custom license
- Last commit
- Sep 2026
- Last release
- May 2026
- Language
- Python
13,120 on 21 Apr 202617,553 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Memori wraps an LLM client (OpenAI, Anthropic, Gemini, Bedrock, DeepSeek, Grok) and records conversations per entity, process and session. It extracts structured memories such as facts, preferences, rules, skills and relationships, and recalls them on later calls. It also ships plugins for OpenClaw and Hermes Agent and a hosted MCP endpoint. The default path uses Memori Cloud with an API key; self-hosting is via a separate BYODB mode.
Who it is for: Developers adding persistent memory to LLM agents and apps
Strengths
- Registers on an existing LLM client; sync, async and streamed calls are covered
- Extracts typed memories: facts, preferences, rules, skills, relationships, events
- Integrations for OpenClaw, Hermes Agent, LangChain, Agno, Pydantic AI and MCP clients
- Reports 87% on LoCoMo at about 721 tokens per query, per its own paper
Weaknesses
- Quickstart depends on the hosted Memori Cloud and an API key
- Advanced Augmentation is a hosted service, rate-limited without an account
- No memories are created unless entity and process attribution is set
- Benchmark figures are vendor-reported, not independently verified
What it needs
- no GPU
- Docker + Compose
- Needs Memori Cloud API key (or BYODB database), LLM provider API key
- Models: OpenAI, Anthropic, Bedrock, DeepSeek, Gemini
Quick answers
Is memori open source?
Its code is public on GitHub under its own license terms; read them before you use it at work.
Can you run memori with Docker?
Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.
Does memori need a GPU?
No, it runs without one.
Also in Memory
See all 20| Rank | Project | Score |
|---|---|---|
| 1 | hindsightAgent memory server with retain, recall and reflect operations | 94 out of 100 |
| 2 | MemPalaceLocal verbatim memory store for AI agents with semantic search and MCP | 90 out of 100 |
| 3 | gbrainLong-term memory store for coding agents, with sources and corrections | 89 out of 100 |
| #4 | Mem0Long-term memory layer that stores and retrieves user facts for LLM agents | 88 out of 100 |
| #5 | SupermemoryMemory and context engine for AI agents, with user profiles and hybrid search | 87 out of 100 |