data-ai MCP Server
Multi-repo semantic code search MCP server in Rust — hybrid vector + BM25 retrieval, tree-sitter AST chunking, fully offline. For OpenCode, Claude Code, Cursor, and any MCP client.
Discovered via github-seeds:mcp-hot and last synced 3mo ago.
Install instructions not detected yet
Check the source repository for the latest setup steps.
Typical alternative
`.c`, `.h`
~hundreds of MB, CPU-only, no Docker
int
`.rs`
string
Override path to `scip-csharp` helper
string
Creates index synchronously before searching (~30-60s)
Automatically create index if it doesn't exist
Notes
Creates minimal placeholder immediately, then indexes in background via incremental refresh
`query`, `limit`, `compact` (default: true), `filter_path`
`minilm-l6`
1024
Without Cache
`symbol`, `limit` (default: 50)
Endpoint
General use
Description
Meaning
`/status`
384
Embedding cache MB (default: 500)
Max matches per file
Index exists but has 0 chunks (placeholder, indexing in progress)
`/search`
768
Log level (e.g. `codesearch=debug`)
Short
Candidates to rerank
Index has chunks and is fully indexed
Behavior
1024
Description
RRF fusion parameter
Error accessing or reading index
ID
Medium
Solution
Creates index synchronously before starting server (~30-60s)
Parameters
384
384
Download
Maximum results
No index database exists
384
Embedding batch size
TreeSitter[Tree-sitter AST] Explore -->
bool
string
int
Manage embedding cache
Technology
string
string
MCP mode: auto, client, local
`.js`, `.jsx`
Versioned via `metadata.json`
`"semantic"` \
string
Path to repos.json
bool
Manage repository groups
Extensions
fastembed + ONNX Runtime (CPU)
bool
Multi-repo routing
string
Always connects to serve, fails if not running
Idle eviction timeout (default: 1800)
string
string
Always uses local DB (classic single-repo stdio)
`.go`
string
`"index"` \
int
`.java`
Type
int
string
Serve mode port (default: 39725)
`.py`
3-layer: in-memory (Moka) → persistent disk → query cache
Tree-sitter AST parsing
`.ts`, `.tsx`