Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_RRF_K | No | RRF k parameter (lower = more weight on top ranks) | 60 |
| MCP_SSE_PORT | No | MCP SSE endpoint port | 8081 |
| MCP_WEB_PORT | No | Web UI port | 8080 |
| MCP_DOCS_PATH | No | Path to .md files inside container | /docs |
| MCP_GIT_TOKEN | No | GitHub token for private repos | |
| MCP_TRANSPORT | No | MCP transport: 'sse' or 'stdio' | sse |
| MCP_GIT_BRANCH | No | Branch to sync | main |
| MCP_GIT_REPO_URL | No | Knowledge repo URL (enables git sync) | |
| MCP_GIT_AUTO_PUSH | No | Auto-commit + push bugfix summaries | true |
| MCP_MIN_RELEVANCE | No | Minimum relevance % to return (0 = no filter) | 0 |
| MCP_CHUNK_STRATEGY | No | Chunking strategy: 'heading', 'fixed', or 'hybrid' | heading |
| MCP_RERANKER_MODEL | No | Reranker model name | cross-encoder/ms-marco-MiniLM-L-6-v2 |
| MCP_WEBHOOK_SECRET | No | GitHub webhook secret (enables /webhook/github HMAC verification) | |
| MCP_EMBEDDING_MODEL | No | Embedding model name | all-MiniLM-L6-v2 |
| MCP_RRF_BM25_WEIGHT | No | BM25 keyword search weight in RRF fusion | 1.0 |
| MCP_RERANKER_ENABLED | No | Enable cross-encoder reranker for higher precision | false |
| MCP_GIT_SYNC_INTERVAL | No | Pull interval in seconds (0 = disabled) | 300 |
| MCP_RRF_VECTOR_WEIGHT | No | Vector search weight in RRF fusion | 1.0 |
| MCP_EMBEDDING_PROVIDER | No | Embedding provider: 'local' (free, private) or 'openai' | local |
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 | |