Agent Ready SDKs

Official client SDKs for the Agent Ready API — scan any public URL for AI agent-readability against the Vercel Agent Readability Spec, the llmstxt.org standard, and agent-protocol manifests (MCP server cards, A2A, agents.json, agent-permissions.json, UCP, x402, NLWeb).

These are libraries you import into your own code. For a terminal tool use the agent-ready-scanner CLI; for MCP-native tools use agent-ready-mcp.

Language	Package	Registry	Source
JavaScript / TypeScript	agent-ready-client	npm	js/
Python	agent-ready-client	PyPI	python/

Both wrap the same REST API (scan / getScan / listScans / ask), are zero-runtime-dependency, and expose the same AgentReady client shape.

// JS / TS
import { AgentReady } from "agent-ready-client";
const ar = new AgentReady({ apiKey: process.env.AGENT_READY_API_KEY });
const scan = await ar.scan("https://example.com");
console.log(scan.vercelScore, scan.vercelRating);

# Python
from agent_ready import AgentReady
ar = AgentReady(api_key="ar_live_...")
scan = ar.scan("https://example.com")
print(scan["vercelScore"], scan["vercelRating"])

Develop

# JS/TS
cd js && npm install && npm test && npm run build

# Python
cd python && python -m pytest tests/   # (pytest), then: python -m build

Publish

• npm: cd js && npm publish (runs prepublishOnly → tsup build).
• PyPI: cd python && python -m build && python -m twine upload dist/*.

Links

• API docs & OpenAPI 3.1 spec: https://agent-ready.dev/docs/api · https://agent-ready.dev/api/v1/openapi.json
• Methodology: https://agent-ready.dev/methodology

License

MIT © Agent Ready