# rohitg00/agentmemory

1 Persistent memory for AI coding agents based on real-world benchmarks

Repository: https://github.com/rohitg00/agentmemory
Canonical: https://ross.abutalabs.com/products/agentmemory
Homepage: https://agent-memory.dev
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: agentmemory, agents, ai, claude, claudecode, codex, copilot, cursor, genai, harness, hermes, memory, openclaw
Last push: 2026-08-24T06:11:16+00:00

## Health v2 (maintenance only)
Score: 81/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 98, longevity 13
- inputs: {"age_days": 189, "days_push": 9, "days_rel": 17, "gap_med": 0.0, "n_releases_24m": 51}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 27526, forks 2361 (observed 2026-08-28T04:11:47.758504+00:00)

## What it is
agentmemory is a local-first persistent memory runtime for AI coding agents like Claude Code, Cursor, Copilot CLI, and any MCP client. It auto-captures session events via hooks, stores them as observations, memories, lessons, and knowledge graphs, and recalls them across sessions using hybrid BM25/vector/graph retrieval with zero external databases.

## Use cases
- give my coding agent persistent memory across sessions
- stop re-explaining project context to claude code every session
- store and recall lessons and preferences for AI coding agents
- add memory to cursor or copilot cli via mcp
- compress agent session history to reduce token usage
- local memory store for AI agents without a vector database
- track what my AI agent did in previous sessions

## When to choose
- you use MCP-compatible coding agents and want durable cross-session memory
- you want a fully local, zero-external-database memory runtime
- you need token savings versus replaying full context each session
- you want auto-capture hooks with no glue code

## When to avoid
- you need a shared multi-user cloud memory service
- you need a general-purpose vector database for your own application embeddings
- you only need simple prompt/context files without retrieval tooling

## Facets
- artifact type: service
- maturity: active
- function: rag, mcp, search-engine, agent-framework, developer-tools
- domain: large-language-models, developer-tools, self-hosted
- platform: cli, cross-platform, self-hosted
- tags: persistent-memory, mcp-server, coding-agents, local-first, memory-runtime, claude-code, hybrid-retrieval, knowledge-graph, ai-agents, nodejs

## Member repositories
- rohitg00/agentmemory (main) score 81

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:47.758504+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:54:31.490790+00:00, confidence not recorded.
  - readme: https://github.com/rohitg00/agentmemory (fetched 2026-08-28T04:11:47.758504+00:00, sha c6054671877a)
  - homepage: https://agent-memory.dev (fetched 2026-08-29T07:51:54.506455+00:00, sha 9180d1b04b04)
  - site_page: https://www.agent-memory.dev/docs (fetched 2026-08-29T07:51:54.516222+00:00, sha 3fbf2b550ef2)
- Data as of 2026-08-30T08:39:29.467469+00:00.