The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the Memory MCP Lite listing page.
A small, opinionated memory server for AI coding assistants (Windsurf, Cursor, Claude Desktop — anything that speaks MCP).
It runs locally, stores durable knowledge on your disk, and tries very hard to stay out of your agent's way until you actually need it.
Most AI clients already have some form of short-term memory. They remember the current conversation, maybe a few rules you've set, and that's about it. What they don't give you is a place to park things that should outlive the session — the architectural decision you made last week, the one weird build command for this repo, the gotcha that bit you three times in a row.
memory-mcp-lite is that place. It stores:
It deliberately does not store raw chat transcripts, replace your client's built-in rules, run embeddings or vector search, or need a server or cloud connection.
Memory lives in a tree:
On top of the tree you can draw optional graph-lite edges between any two nodes — related_to, depends_on, affects, caused_by, supersedes, references. Handy when one decision obsoletes another, or a gotcha only matters in the context of a specific command.
The retrieval side is built to be cheap. The server's instructions push agents through three stages, from least to most expensive:
In practice this means your agent asks for a summary first, and only pays for the big payload when it has a specific reason to. If you skip this policy, you just end up dumping a bunch of stringly-typed JSON into context for no reason.
@libsql/client)@modelcontextprotocol/sdk)You can point it at a local file, a remote libSQL instance, or a Turso database — they all work the same.
The fast path is to let your MCP client fetch the package via npx.
Windsurf — ~/.codeium/windsurf/mcp_config.json:
Claude Desktop — ~/Library/Application Support/Claude/claude_desktop_config.json:
Same pattern for any other MCP-compatible client; only the config file path changes.
Then point your client at the compiled bundle:
If you want to iterate on the code without a build step, tsx works:
By default: ~/.memory-mcp/memory.db. Override it with any of:
| Env var | Purpose |
|---|---|
MEMORY_DB_PATH | Full path or libsql://… / file: URL. |
MEMORY_DATA_DIR | Directory; the file is still memory.db. |
DATABASE_URL | Accepted for backwards compatibility. |
MEMORY_DB_AUTH_TOKEN | Bearer token for remote libSQL / Turso. |
So running against Turso is just:
Nine tools, all returning both a human-readable JSON block and a structuredContent object for programmatic clients. The server also ships a strict description and annotations payload for each tool so agents can pick the right one without guessing.
| Tool | Reach for it when… |
|---|---|
get_global_summary | recurring preferences, cross-project conventions |
get_project_summary | architecture, key decisions, long-term project context |
get_task_summary | resuming a specific piece of work |
search_memory_light | summaries aren't enough; you want compact candidates |
get_memory_detail | you've picked a candidate and need the full body |
remember_decision | an architecture choice, trade-off, or rejected path |
remember_fact | a command, env note, gotcha, link, or convention |
upsert_project_summary | after an arch change or new convention worth recording |
upsert_task_summary | after progress, blockers, or a plan change |
The retrieval discipline the server asks agents to follow:
Projects are looked up in this priority order:
This means the same project keeps the same memory even if different clients hand you slightly different paths, and moving a repo doesn't orphan everything you've stored.
The schema is defined in apps/server/src/db/schema.ts and re-asserted on every startup by ensureSchema() (see apps/server/src/db/migrate.ts). That function is also where the FTS5 virtual table and its triggers get created — Drizzle doesn't manage virtual tables, so we do it ourselves with plain SQL. It's idempotent, so there's nothing to run manually.