semantic-memory-mcp
by guangbin79
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MEMORY_HOST | No | Bind address | 127.0.0.1 |
| MEMORY_PORT | No | Bind port; port 0 = OS-assigned random port | 8765 |
| MEMORY_RRF_K | No | RRF fusion constant | 60 |
| MEMORY_TOKEN | No | REST sidecar Bearer auth; required when binding non-loopback | |
| MEMORY_DEVICE | No | Embed device: cuda (default, auto-falls back to CPU) / cpu | cuda |
| MEMORY_DB_PATH | No | Database path | ~/.local/share/memory-mcp/memory.db |
| MEMORY_DEDUP_SIM | No | Cosine threshold for vector near-duplicate ingest dedup | 0.98 |
| MEMORY_MODEL_PATH | No | GGUF model path; empty = BM25-only fallback | |
| MEMORY_INJECT_BOOST | No | Telemetry feedback: score × (1 + α·ln(1+injections)); 0 = off | 0.15 |
| MEMORY_DISCOVERY_FILE | No | Service-discovery file; off = disabled | <db>.server.json |
| MEMORY_RECALL_MIN_COSINE | No | Cosine floor for vector-only hits (BM25 hits always pass); 0 = off | 0.45 |
| MEMORY_TTL_ARTIFACT_DAYS | No | TTL for artifact kind; 0 = never expire | 90 |
| MEMORY_TTL_INJECTION_DAYS | No | TTL for injection telemetry; 0 = never expire | 90 |
| MEMORY_MAINTENANCE_ON_START | No | Run the maintenance routine at service startup | 1 |
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
Server capabilities have not been inspected yet.
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
No tools | |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
This server cannot be deployed
Maintenance
ActivityMaintained
ResponsivenessNo issues