Research Crew
Research Crew is a team of four AI agents, a planner, a researcher, a critic and a writer, that produce a short cited brief on a market you pick. You watch each step as it happens, including what each agent decided and what the run costs. Briefs export to PDF or Markdown, and a market the sources do not cover ends with an honest "not enough evidence" report.
Measured numbers
- Tests
- 385
- Browser tests
- 11
- Lighthouse performance
- 95
- Accessibility
- 100
- Best practices
- 100
- SEO
- 100
Home page, mobile. Copied from the showcase's own README on 24 October 2026.
Open the live site (opens in a new tab) crew.phoenixcodes.org
On a phone
At desktop width
What it proves
- Multi-agent orchestration as an explicit state machine, with a critic that rejects unsupported claims.
- Every claim in the brief links to its source.
- Cost and token use are visible for each run.
- Search sits behind an adapter, so the built-in sample sources can be replaced by live web search through Tavily.
How this adapts to your project
Give the crew your sources, whether a document store, a research feed or live web search, and your house format for briefs. The critic's rules decide what counts as supported, so a legal, financial or competitive team can set its own bar.