ai-memory

thedotmack/claude-mem

A shared memory layer so AI coding agents remember past sessions — claude-mem captures what an AI agent does in a session, compresses it with AI, and re-injects relevant context into future sessions using ChromaDB/SQLite embeddings. It's built as a plugin for Claude Code, OpenClaw, Codex, Gemini, Copilot and others, aiming to give every agent persistent, cross-session memory.

vectorize-io/hindsight

Agent memory that learns instead of just recalling history — Hindsight is an agent memory system designed to help AI agents learn over time rather than simply retrieve past conversations. It claims to outperform RAG and knowledge-graph approaches on long-term memory benchmarks, backed by a published paper. Useful for developers building agents, coding assistants, or AI products that need persistent, evolving memory via server, embedded Python, or MCP integration.

topoteretes/cognee

Persistent long-term memory for AI agents, free and open source — Cognee gives AI agents long-term memory using small models, combining vector and graph databases for context engineering instead of fine-tuning. It targets developers building agents that need to recall knowledge across sessions, with community plugins extending its core.

supermemoryai/supermemory

A benchmark-leading memory layer built for self-hosted AI agents — Supermemory ranks #1 on LongMemEval, LoCoMo, and ConvoMem benchmarks, claiming 95% Recall@15 with 99.4% context reduction and ~50ms user profiles. Built on Cloudflare Workers, Drizzle, and Postgres, it offers an npm/pip API plus self-hosting, aimed at developers building AI agents that need fast, persistent memory.

screenpipe/screenpipe

Screenpipe records your work screen to surface automation targets — Screenpipe continuously captures screen and audio on company computers, then maps workflows to find tasks worth automating and feeds context to AI agents. Backed by YC S26, it's built local-first and privacy-conscious, aimed at teams wanting to understand and automate real work patterns rather than guess at them.

plastic-labs/honcho

Memory library for building stateful agents

felinics/Memoh

Open-source cloud computers that keep AI agents running 24/7 — Memoh gives each AI agent its own always-on cloud workspace with filesystem, browser, network and long-term memory. Bring your own API key or host Claude Code/Codex agents, then chat via Telegram, Discord, Lark, WeChat or Web UI across sessions.

vshulcz/deja-vu

A local memory layer that mines your coding agents' own session logs — Deja-vu indexes months of session history already saved on disk by Claude Code, Codex, Cursor and 39 other agents, then surfaces the relevant fix to whichever agent runs next. It works via local search, MCP and hooks with no LLM calls, shipped as a single Go binary, and reports 97.2% R@5 on LongMemEval-S.

HelpCode-ai/anythingmcp

Self-hosted gateway that wraps 517 APIs as MCP tools — AnythingMCP connects REST/OpenAPI, SOAP, GraphQL, OData and SQL systems to Claude, ChatGPT and Copilot without writing a custom MCP server. It ships 517 ready-made connectors, including SAP, Etsy and Amazon Seller, with 23 needing no API key, plus RBAC, SSO and audit logs for enterprise use.

aru-labs/lossless-memory

Lossless long-term memory for a personal AI: never summarize, keep every line, put a timestamp on everything.

Avinash-jetwani/jevmem

Persistent memory for Claude Code, Cursor and Codex sessions — jevmem auto-saves decisions, rules and failed approaches from AI coding chats into JEVMEM.md and resurfaces them next session, checking commands and edits against saved rules first. It now works with Cursor, Codex and Claude Desktop via MCP, with test data showing fewer repeated mistakes and forbidden changes.

mrveiss/AutoBot-AI

Self-hosted, agentic AI you own — voice, vision-driven browser & computer control, visual workflows, human-in-the-loop approvals, multi-user RBAC, and a knowledge graph. Runs on your infrastructure. Your data. Your AI.