mcp-memory
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_MEMORY_DIR | No | Directory for memory data files. | ./data |
| EMBEDDING_MODEL | No | Embedding model name. | |
| EMBEDDING_API_KEY | No | API key for the embedding service. | |
| EMBEDDING_API_URL | No | OpenAI-compatible embedding API base URL. If not set, semantic search is disabled. | |
| SIMILARITY_THRESHOLD | No | Minimum cosine similarity to return results. | 0.45 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| write_memoryB | Write a new memory. Choose a category that fits the content. |
| read_memoriesA | Read memories with pagination (newest first). Use search_memories for keyword lookup. |
| search_memoriesA | Search memories by keyword (exact substring match). For semantic search, use recall with op:similar or op:fuse. |
| update_memoryC | Update an existing memory by ID. |
| delete_memoryC | Delete a memory by ID. |
| get_statsB | Get memory statistics. |
| recallA | Multi-mode memory retrieval. • day: Fetch memories by date (date or from+to, YYYY-MM-DD) • timemachine: See what happened N days/months/years ago today • similar: Semantic recall — find related memories even with different wording (requires embedding API) • fuse: ★ 3-way fusion (BM25 + semantic + tag graph → RRF). Best for finding specific things. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 7 tools
The core CRUD tools (write/read/update/delete) are clearly distinct. However, search_memories and recall overlap in retrieval, though recall's modes (similar/fuse) are semantically distinct from exact substring matching, so boundaries are mostly clear.
All tool names follow a consistent verb_noun pattern in snake_case (write_memory, read_memories, search_memories, update_memory, delete_memory, get_stats). 'recall' is a single verb but fits the action-oriented style without breaking consistency.
Seven tools is well-scoped for a memory management server, covering all essential operations without redundancy or bloat. Each tool serves a clear purpose and earns its place.
The set provides full CRUD lifecycle (create, read, update, delete), plus dedicated search and statistical functions. The multi-mode recall covers advanced retrieval needs, leaving no obvious gaps in the memory domain.