Question 1

What is Arlong?

Accepted Answer

Arlong is a web search and research API for AI agents, available through REST and MCP. It retrieves sources, screens untrusted content and returns evidence for the agent’s task.

Question 2

How does Arlong handle source verification and claim corroboration?

Accepted Answer

The research pipeline gathers source evidence, checks coverage and looks for supporting or conflicting evidence. A source link alone does not verify a claim, and syndicated copies should not count as independent confirmation. Review the returned evidence and coverage before relying on an answer.

Question 3

Can I use Arlong Screen with another search API?

Accepted Answer

Yes. Arlong Screen works independently of Arlong search. Submit text, HTML or a supported public URL through the API or MCP tool, and use only the permitted safe_content from an allowed full-mode result.

Question 4

Does screening guarantee that prompt injection cannot reach my agent?

Accepted Answer

No. Screening is a protective layer with measured limitations. Allowed content remains untrusted evidence. Limit agent permissions and require authorization for sensitive actions.

Question 5

Where can I compare search APIs for agents?

Accepted Answer

The Arlong blog includes guides to source verification, Exa versus Tavily, and Tavily alternatives. The guides cite official documentation and separate documented capabilities from engineering recommendations.

Give your AI agents access to the web, with evidence they can inspect and prompt injection screening before content reaches their context.