Qveris Coding

QVeris — External API Code Generation

When a coding project needs to call external APIs or third-party tools (weather, stocks, search, currency, data retrieval, etc.), follow this workflow to discover a working API and generate production code.

Step 1 — Search & Verify via MCP

Use the QVeris MCP server to find and test tools.

1. Search: Call the search_tools MCP tool with a query that describes the capability you need (not specific parameters).
   • Good query: "current weather data", "stock price API", "currency exchange rates"
   • Bad query: "get weather for London" (too specific)
2. Test: Call execute_tool with the tool_id from search results and pass parameters via params_to_tool (JSON string). Refer to examples returned by search if available.
3. You may call multiple MCP tools in a single response. Iterate between search_tools and execute_tool until you find a tool that works.

Step 2 — Generate Production Code

Once you have confirmed a working tool_id from Step 1, generate code that calls the QVeris REST API directly. Do not use the MCP tool results as the final output — generate real, standalone code.

Rules for generated code:

• Use the verified tool_id directly (no search call needed in production code).
• Set HTTP request timeout to 5 seconds.
• Handle error responses correctly (non-200 status, success: false).
• Place the QVERIS_API_KEY in the code (read from environment variable or config).

QVeris REST API Reference

Base URL: https://qveris.ai/api/v1

Authentication: Bearer token in the Authorization header.

Authorization: Bearer YOUR_QVERIS_API_KEY

Execute Tool Endpoint:

POST /tools/execute?tool_id={tool_id}

Request body:

{
  "search_id": "string",
  "session_id": "string",
  "parameters": {
    "city": "London",
    "units": "metric"
  },
  "max_response_size": 20480
}

Response (200 OK):

{
  "execution_id": "string",
  "result": {
    "data": { ... }
  },
  "success": true,
  "error_message": null,
  "elapsed_time_ms": 847
}

Fallback — CLI Script (if MCP is unavailable)

If MCP tools are not available or not working, use the CLI script instead.

Search:

uv run {baseDir}/../scripts/qveris_tool.py search "weather forecast API" --limit 5

Execute (to verify a tool works):

uv run {baseDir}/../scripts/qveris_tool.py execute <tool_id> --search-id <id> --params '{"city": "London", "units": "metric"}'

After verifying a working tool via CLI, proceed to Step 2 above to generate production code using the REST API.