Agent Memory & State

Platform for building stateful agents with advanced memory that learn and self-improve over time.

Universal memory layer for AI agents. Persistent, multi-session memory across models and environments.

Open-source memory infrastructure for AI agents, providing persistent memory and context management for building stateful, long-running agents.

Open-source memory service for agents with benchmarked retrieval, structured entities, and API/SDK support for persistent personalized memory in tooling workflows.

MCP server for persistent AI agent memory with atomic notes, semantic linking, and SQLite or PostgreSQL storage.

State-of-the-art long-term memory for AI agents by Vectorize. Fully self-hosted, MIT-licensed, with integrations for LangChain, CrewAI, LlamaIndex, Vercel AI SDK, and more.

Memory API and app for AI agents that provides fast, scalable, context-aware storage and retrieval across projects. MIT licensed.

Fully local long-term memory layer for AI agents with a four-tier progressive pipeline and zero external dependencies. MIT licensed.

AI memory platform for agents that builds a self-hosted knowledge graph for persistent, long-term memory across sessions. Apache 2.0 licensed.

Lightweight state kernel and local control plane for long-running AI agent loops with goal tracking, auto-wake, and verifiable handoffs. MIT licensed.

Local-first context and memory engine that keeps a billion-token encoded pool on disk and pages relevant slices into a small resident window for Ollama and other local LLMs. Apache 2.0 licensed.

Cross-client AI memory hub where agents share one physical SQLite database via file-level pointers. Zero cloud, bi-temporal governance with supersession chains, hash-chain audit, and MCP. Apache-2.0.

Memory layer that retires a corrected fact by key and can undo the correction later from the key alone. Zero-dependency core with an MCP server.

MIT-licensed MCP server for agent memory, backed by a hosted engine that re-ranks recall from reported outcomes rather than similarity alone, with one key or OAuth across Claude Code, Cursor, VS Code, and ChatGPT.

Open-source memory runtime for AI agents that compiles reproducible, provenance-tagged context bundles instead of relying on query-time retrieval, self-hosted on Postgres and pgvector with Python and TypeScript SDKs. Apache 2.0 licensed.

Causal proof engine and operating system for AI agents built on Strata, an append-only signed log, with no vectors and no RAG, where every answer carries its proof as a memory id, an edge path or a receipt and a failure is traced back to the commit that caused it. AGPL 3.0 licensed.

Verifiable terminal coding agent combining an append-only evidence hash chain, a single deterministic verification gate, and a manual-verification checklist for what it could not check (multi-provider, MCP-ready, MIT).