Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@agent-memory search my memory for notes about the payment integration"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Local markdown memory for coding agents — any provider, any tool.
Without a shared layout, each vendor owns a graveyard — session archives that never become durable identity unless you re-teach every new agent. See abi/WHY.md for the full rationale (vs RAG, vs vaults, vs vendor memory). Reference-implementation roadmap: ROADMAP.md.
Your memory lives in folders you own: ~/.agents/memory (identity + project map) and <repo>/.agents/memory (per-project work). Your Agent — or any host that reads markdown and optional MCP — uses the same layout.
Markdown is the source of truth. The Python MCP in this repo is the reference implementation, not a proprietary backend.
Shipped under abi/ — versioned, documented, implementation-agnostic:
Doc | What |
Overview + conformance | |
Why this exists (vs RAG, vaults, vendor graves) | |
Reference-implementation roadmap (not ABI) | |
Folder taxonomy | |
Where facts go | |
Tool contract | |
AGENTS / CLAUDE binding | |
Windows, symlinks, per-machine paths | |
Sync, injection, scan.json, | |
Catalog / extract / distill pipeline |
Install copies abi/LAYOUT.md → ~/.agents/memory/LAYOUT.md. Injection and CLI flags: abi/INSTALL.md. Machine-readable: python -m agent_memory --help-json.
Related MCP server: mcp-chest-memory
Python lives in src/agent_memory/ (pip install -e . from the clone).
Piece | Role |
| MCP: search / add / register / inventory / sync / ingest |
| Layout engine, taxonomy, injection |
| Canonical |
| Disk vs |
| Chat catalog + extract pipeline |
| Agent install skill |
Templates only in git: examples/*.example.*. Live store: ~/.agents/memory. Run python -m agent_memory consolidate if markdown leaked into this clone.
Coding agents: follow AGENTS.md for install.
Path | What |
| Always-on identity |
| slug / path / role / stack / status |
| Cross-cutting typed memory (created on first write) |
| Link to real tree (not a copy) |
| Files only when used — staging inbox, then typed paths on write |
| Short inject; |
Search unions user + all registered project trees. Chat bodies stay in product folders; chats-index.md is titles + paths only.
Same interpreter for pip and MCP:
python -m pip install -e .
python -m agent_memory sync --init
python -m agent_memory inventoryFill ~/.agents/memory/USER.md and scan.json (from examples/*.example.* on first run). Reload your Agent after MCP merge.
See abi/PLATFORM.md for Windows symlink stubs and hardlink fallback.
python -m agent_memory inventory
python -m agent_memory inventory --json
python -m agent_memory inventory --register SLUG "/path/to/repo" "role" "stack"
python -m agent_memory sync
python -m agent_memory --help-json
python -m agent_memory sync --help-json
python -m agent_memory inventory --help-json
python -m agent_memory ingest catalogAfter install, agent-memory on PATH is the same as python -m agent_memory.
This server cannot be installed
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Shared, governed long-term memory for AI agents across tools and sessions via MCP and REST.
Your memory, everywhere AI goes. Build knowledge once, access it via MCP anywhere.
Shared long-term memory vault for AI agents with 20 MCP tools.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Lolaplex/agent-memory'
If you have feedback or need assistance with the MCP directory API, please join our Discord server