Developer Automation Recipe

Use this recipe to discover and test a developer-facing API, such as repository metadata, package metadata, release lookup, or issue search.

Quickstart#

export QVERIS_API_KEY="sk-..."
qveris init --query "GitHub repository metadata API" --params '{"owner":"QVerisAI","repo":"qveris-agent-toolkit"}' --json

CLI#

qveris init \
  --query "GitHub repository metadata API" \
  --params '{"owner":"QVerisAI","repo":"qveris-agent-toolkit"}' \
  --max-size 20480 \
  --json

Audit the call:

qveris usage --execution-id "exec_..." --json
qveris ledger --limit 5 --json

Python SDK#

import asyncio
from qveris import QverisClient
 
async def main() -> None:
    client = QverisClient()
    try:
        discovered = await client.discover("GitHub repository metadata API", limit=5)
        if not discovered.results:
            print("No capabilities found.")
            return
        tool = discovered.results[0]
        inspected = await client.inspect(tool.tool_id, search_id=discovered.search_id)
        selected = inspected.results[0] if inspected.results else tool
        result = await client.call(
            selected.tool_id,
            {"owner": "QVerisAI", "repo": "qveris-agent-toolkit"},
            search_id=discovered.search_id,
        )
        print(result.model_dump())
    finally:
        await client.close()
 
asyncio.run(main())