Repolode
Submit
MemoriLabs

MemoriLabs/Memori

Agent-native memory infrastructure that turns LLM agent conversations and execution into persistent, structured memory for Python and TypeScript apps.

Open-source alternative to:Mem0Zep

View on GitHubWebsite
Stars
17,149
Forks
3,852
Open issues
41
Last commit
18 days ago

What is it?

Memori is a memory layer for LLM agents. It captures what an agent says and does, including tool calls, decisions and outcomes, and stores them as structured state that later requests can recall. It is LLM-agnostic and works alongside the datastores and frameworks an application already uses.

The SDKs wrap an existing OpenAI client or similar LLM client so conversations are recorded and recalled in the background, without changing prompts. Interactions are grouped by entity (such as a user), process (such as an agent) and session. Advanced Augmentation extracts facts, preferences, rules, skills and relationships from those interactions.

Beyond the SDKs, Memori provides an MCP endpoint for clients such as Claude Code and Cursor, plugins for OpenClaw and Hermes Agent, and a bring-your-own-database option. It can be deployed as managed cloud, single-tenant cloud, VPC or on-premises.

Key features

  • 01Records LLM conversations and agent execution automatically in the background.
  • 02Supports OpenAI, Anthropic, Gemini, Bedrock, DeepSeek and Grok clients.
  • 03Integrates with LangChain, Agno and Pydantic AI.
  • 04Connects to Claude Code, Cursor and other MCP clients with one command.
  • 05Organizes memories by entity, process and session.
  • 06Offers TypeScript and Python SDKs.

Use cases

  • Giving a customer support agent memory of each user's preferences across sessions.
  • Letting coding agents retain project conventions and past decisions.
  • Reducing redundant tool calls in production agents by reusing recorded execution traces.
  • Deploying agent memory inside a private VPC or on-premises environment.

Quick start

  1. 1

    Get a Memori API key at app.memorilabs.ai

  2. 2

    Install the Python SDK

    pip install memori
  3. 3

    Set MEMORI_API_KEY and OPENAI_API_KEY in your environment, then run your LLM code with Memori registered

Things to know

  • Memori Cloud requires a Memori API key, and the LLM provider needs its own API key.
  • Advanced Augmentation is rate-limited without an account.
  • Memories are only created when interactions are attributed to an entity and process.

Topics

#agent-memory#llm#ai-memory#python#typescript#rag#mcp#agent#agenticai#ai

FAQ

What is Memori?
Agent-native memory infrastructure that turns LLM agent conversations and execution into persistent, structured memory for Python and TypeScript apps. Memori is a memory layer for LLM agents. It captures what an agent says and does, including tool calls, decisions and outcomes, and stores them as structured state that later requests can recall. It is LLM-agnostic and works alongside the datastores and frameworks an application already uses.
Is Memori free to use? What license does it have?
Memori is open source under the Other license. Check the license terms before using it in commercial products.
Is Memori still maintained?
Actively maintained. The last commit to Memori was on Oct 3, 2026.
What can I use Memori for?
Giving a customer support agent memory of each user's preferences across sessions. Letting coding agents retain project conventions and past decisions. Reducing redundant tool calls in production agents by reusing recorded execution traces. Deploying agent memory inside a private VPC or on-premises environment.
How do I install Memori?
Get a Memori API key at app.memorilabs.ai Install the Python SDK: pip install memori Set MEMORI_API_KEY and OPENAI_API_KEY in your environment, then run your LLM code with Memori registered
Where can I get Memori?
The source code and installation instructions are on GitHub: https://github.com/MemoriLabs/Memori

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects