Quick Start

This guide assumes you have already installed Kodit and have verified it is running.

1. View the API Docs

Open: http://localhost:8000/docs (replace with your own server URL)

2. Index a Repository

Replace the repository URL with the one you want to index. This is a small, toy application that should index quickly.

curl http://localhost:8080/api/v1/repositories \
-X POST \
-H "Content-Type: application/json" \
-d '{
  "data": {
    "type": "repository",
    "attributes": {
      "url": "https://gist.github.com/philwinder/7aa38185e20433c04c533f2b28f4e217.git"
    }
  }
}'

3. Check the Status of the Indexing Process

curl http://localhost:8080/api/v1/repositories/1/status

Wait for the indexing process to complete. If you see any errors at this stage, check your logs and consult the troubleshooting guide.

4. Search for Code

When indexing is complete, you can search through the index of the repository:

curl http://localhost:8080/api/v1/search \
-X POST \
-H "Content-Type: application/json" \
-d '{
  "data": {
    "type": "search",
    "attributes": {
      "keywords": [
        "orders"
      ],
      "code": "func (s *OrderService) GetAllOrders() []Order {",
      "text": "code to get all orders",
    }
  }
}'

Check the API docs for more endpoints and examples.

5. Integrate with your AI coding assistant

Now add the Kodit MCP server to your AI coding assistant.