Skip to main content
Glama
adam-hanna

semantic-search-mcp

by adam-hanna

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
SEMANTIC_SEARCH_DB_PATHNoIndex database location.semantic-search/index.db
SEMANTIC_SEARCH_MIN_SCORENoMinimum relevance threshold (0-1)0.3
SEMANTIC_SEARCH_BATCH_SIZENoFiles per batch50
SEMANTIC_SEARCH_DEBOUNCE_MSNoFile watcher debounce in milliseconds1000
SEMANTIC_SEARCH_USE_QUANTIZEDNoUse INT8 quantized model (30-40% faster)true
SEMANTIC_SEARCH_EMBEDDING_MODELNoEmbedding modeljinaai/jina-embeddings-v2-base-code
SEMANTIC_SEARCH_MAX_FILE_SIZE_KBNoSkip files larger than this (KB)512
SEMANTIC_SEARCH_EMBEDDING_THREADSNoONNX runtime threads4
SEMANTIC_SEARCH_EMBEDDING_BATCH_SIZENoTexts per embedding call8

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

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
initializeB

Initialize or re-initialize the semantic code search system.

Loads the embedding model and builds or updates the code index. Use force_reindex=True to rebuild the entire index.

Progress will be reported during indexing.

search_codeA

Search the codebase using semantic similarity.

Use natural language descriptions of code you're looking for:

  • "function that handles user authentication"

  • "error handling for HTTP requests"

  • "database connection initialization"

  • "unit tests for the payment service"

Returns ranked code snippets with file locations and relevance scores. Combines vector similarity with keyword search for best results.

Best for: Finding functions/classes by purpose or behavior. Not for: Exact pattern matching or finding all occurrences of a variable. Use Grep for exact patterns; use semantic search to find relevant files first.

reindex_fileA

Re-index a specific file for search.

Use when a file has been modified but not yet re-indexed, or when you want to force a refresh of a file's embeddings.

pause_watcherA

Pause the file watcher.

Events that occur while paused are discarded. Use resume_watcher to start watching again.

resume_watcherA

Resume the file watcher after pausing.

Starts watching for file changes again.

cancel_indexingA

Cancel any running indexing job.

The indexing will stop after the current file completes. Partial results are kept in the index.

clear_indexA

Clear all indexed data.

Removes all files and chunks from the index. The index will be empty until reindex is called.

exclude_pathsA

Add paths to exclude from indexing (session-only).

Patterns use glob syntax. Examples:

  • "node_modules" - exclude any path containing node_modules

  • "*.test.py" - exclude files ending in .test.py

  • "vendor/**" - exclude everything under vendor/

Exclusions reset when the server restarts.

include_pathsA

Remove paths from the exclusion list.

Reverses the effect of exclude_paths for the specified patterns.

get_statusA

Get comprehensive server status.

Returns server state, watcher status, indexing progress, index statistics, and current exclusion patterns.

reindexA

Start a full reindex of the codebase.

Runs in the background - use get_status to monitor progress. Use cancel_indexing to abort if needed.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription
get_status_resourceCurrent index status and statistics.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/adam-hanna/semantic-search-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server