Deep Research API for AI agents

Why agents need a research API

Agents often need multi-source investigation with citations, not just another model completion. ResearchAO exposes research as HTTP endpoints with personal API keys, usage metering, and optional webhooks.

Recommended pattern

1. Create a key in the app (examples use rao_live_...)
2. POST a research run with query + effort via the Deep Research API
3. Store the run id; poll GET or wait for a webhook
4. Pass the structured report (or a summary) back into the agent context

Schedules vs agent loops

If the question is stable, prefer a schedule and let the agent only handle deltas. That reduces prompt churn and credit waste.

What we do not claim

This guide does not claim Model Context Protocol (MCP) support or official plugins for any specific agent framework. Integrate with standard HTTP and the API docs / OpenAPI spec.

Related

• Research for AI Agents
• Pricing