A Mem Mcp
A self-evolving memory layer that turns your AI agent's knowledge into a growing graph rather than a flat store. Exposes eight MCP tools including add_memory_note, search_memories_agentic for connection-aware search, and read_memory_note for detailed retrieval. When you add knowledge, it automatically extracts keywords and tags via LLM, finds semantic neighbors in ChromaDB, then decides whether to link or strengthen existing connections. Memory lives per-project by default or globally if you point CHROMA_DB_PATH elsewhere. Supports OpenAI, Ollama, OpenRouter, and SGLang backends. Tested with Claude Code and includes session hooks to remind the agent to persist learnings. Reach for this when you want your agent to remember context across sessions and have that memory actually organize itself over time.
- GitHub stars
- 34
- tools
- 8
- auth
- Secret needed
- transport
- STDIO
- env vars
- 5
- last updated
- Jan 9, 2026
Tools
8 toolsPublic tool metadata for what this MCP can expose to an agent.
add_memory_noteStore new knowledge (async, returns immediately)
Store new knowledge (async, returns immediately)
No parameter schema in public metadata yet.
search_memoriesSemantic search across all memories
Semantic search across all memories
No parameter schema in public metadata yet.
search_memories_agenticSearch + follow graph connections
Search + follow graph connections
No parameter schema in public metadata yet.
search_memories_by_timeSearch within a time range
Search within a time range
No parameter schema in public metadata yet.
read_memory_noteGet full details (supports bulk reads)
Get full details (supports bulk reads)
No parameter schema in public metadata yet.
update_memory_noteModify existing memory
Modify existing memory
No parameter schema in public metadata yet.
delete_memory_noteRemove a memory
Remove a memory
No parameter schema in public metadata yet.
check_task_statusCheck async task completion
Check async task completion
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageConfiguration
LLM_BACKENDdefault: openaiLLM backend to use (openai, ollama, sglang, or openrouter)
LLM_MODELdefault: gpt-4o-miniLLM model name (e.g., gpt-4o-mini, llama2, etc.)
OPENAI_API_KEY*secretOpenAI API key (required if LLM_BACKEND=openai)
EMBEDDING_MODELdefault: all-MiniLM-L6-v2Sentence transformer model for embeddings
CHROMA_DB_PATHdefault: ./chroma_dbChromaDB storage directory path
Readme
A-MEM: Self-evolving memory for coding agents
mcp-name: io.github.DiaaAj/a-mem-mcp
A-MEM is a self-evolving memory system for coding agents. Unlike simple vector stores, A-MEM automatically organizes knowledge into a Zettelkasten-style graph with dynamic relationships. Memories don't just get stored—they evolve and connect over time.
Currently tested with Claude Code. Support for other MCP-compatible agents is planned.
Quick Start
Install
pip install a-mem
Add to Claude Code
claude mcp add a-mem -s user -- a-mem-mcp \
-e LLM_BACKEND=openai \
-e LLM_MODEL=gpt-4o-mini \
-e OPENAI_API_KEY=sk-...
Details
- Repository
- github.com/diaaaj/a-mem-mcp
- Categories
- AI & LLM Tools, Documents & Knowledge
- Package
a-mem- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active
Related AI & LLM Tools MCP servers
View all
Enquire Mcpoomkapwn/enquire-mcpLong-term memory for AI agents — your Obsidian vault as searchable persistent memory.34stars
Selvageselvage-lab/selvageAn LLM-based code review MCP server with AST-powered smart context extraction33stars
Workflows Mcp Servercyanheads/workflows-mcp-serverStore, query, and create YAML workflow playbooks for LLM agents.32stars
ThumbGate — PreToolUse Agent Firewalligorganapolsky/thumbgateSub-ms PreToolUse firewall, feedback memory, and single-use permits for AI coding agents.31stars
Cuba Memorysleandropg19/cuba-memorysPersistent memory MCP server. 28 tools, BM25+MMR+OOD retrieval, tamper-evident audit, code graph.30stars
Humane Proxyvishisht16/humane-proxyAI safety middleware — detects self-harm and criminal intent in LLM prompts.28stars