Installing Kodit

Unlike many MCP tools, Kodit is designed to run as a server. You’re welcome to use the public server, but if you want to index your own private repositories then you should deploy your own server.

Kodit is a Go binary that hosts a REST API and an MCP server. Although most people deploy Kodit remotely in a container, you can also run it locally.

Docker

docker run -it --rm -p 8080:8080 registry.helix.ml/helix/kodit:latest

Always replace latest with a specific version.

Pre-built Binaries

Pre-built binaries for Linux and macOS are available on the GitHub releases page. Download the appropriate binary for your platform, make it executable, and run it:

chmod +x kodit
./kodit serve

Enrichment Model

Kodit includes a built-in embedding model for semantic search, so basic keyword and semantic search works out of the box. However, to unlock AI-powered enrichments (architecture docs, API docs, database schemas, cookbook entries, and commit summaries), you need to configure an enrichment endpoint that points to an LLM provider.

Without an enrichment model, Kodit will index and search code but will not generate any AI summaries or documentation.

Set the following environment variables to configure an enrichment provider:

ENRICHMENT_ENDPOINT_BASE_URL=https://app.helix.ml/v1
ENRICHMENT_ENDPOINT_MODEL=Qwen/Qwen3-8B
ENRICHMENT_ENDPOINT_API_KEY=your-api-key

See the configuration reference for more provider examples including Ollama, Azure OpenAI, and other LiteLLM-compatible services.

Next Steps

See the deployment guide for Docker Compose and Kubernetes deployment instructions, or jump straight to the quick start to index your first repository.