Research and intelligence agents

Agents that gather evidence from the web and documents, and return sourced findings.

19 agents, each with its inputs, outputs and review level checked by a person.

Research agents gather evidence and return findings you can check. They search the web and documents, read many sources, compare them, and hand back a report, a brief or a structured dataset with the sources attached. The listings distinguish agents that research a market or a competitor from agents that answer a scientific question from papers, and agents that browse sites to complete lookups from agents that only read. What they share is the contract: a question or brief in, sourced findings out, gaps marked rather than filled.

When choosing, look at the sources each agent can reach and at how it records them. For buyer and market research, agents that read forums, reviews and competitor pages matter. For academic work, agents grounded in paper databases matter. For anything that feeds a decision, prefer agents whose acceptance section says evidence is logged and unknowns stay unknown.

The agents

Questions

How is a research agent different from a search engine?

A search engine returns links. A research agent reads the pages, synthesises them into an answer or a report, and cites what it used. The listings record what each agent returns and how it logs sources.

Can research agents make things up?

Any generated text can. The agents here that record sources and mark gaps make that easier to catch, and their acceptance checks say so. The catalogue lists what the vendor documents, not a guarantee of accuracy.

Which agents can act on websites rather than just read them?

Browser agents can fill forms, click through flows and download files. They are listed under the use case web browsing tasks, and their listing names what they can do.