Show HN: Showboat and Rodney, so agents can demo what they've built

Showboat for AI demos drops — fans cheer, skeptics yell “just prompt it”

TLDR: Developer Simon Willison released Showboat and Rodney to let AI helpers create clear, proof-of-work demos, aiming to make “does it work?” obvious. The crowd split between fans who’ll use them as task gates and skeptics saying it’s just a prompt away, with side debates over documentation formats and his switch to Go.

Simon Willison just launched two tools — Showboat and Rodney — to help AI “coding agents” prove what they’ve actually built. In plain English: Showboat creates a step‑by‑step demo document (commands, outputs, even images) so you can see the software work, and Rodney is a helper tool for fetching things from the web. It’s Simon’s cheaper alternative to pricey robot testers and “no human reviews” pipelines, and the comments immediately lit up.

On the excited side, users say this tackles a real pain: getting bots to show working software, especially for websites, is messy. As one supporter put it, closing the loop on a web interface isn’t as simple as a terminal — and they’re eager to try it. Others tossed in “we already have tools” energy, pointing to Microsoft’s Playwright CLI to say, essentially, mind the duplicates.

Then came the spicy counter‑takes: a skeptic argued this could be one prompt away with big AI models like Claude, calling it a skill, not a tool. Documentation purists dove straight into a format war, asking why it uses Markdown instead of more powerful doc formats — and nitpicking the code block syntax. Meanwhile, workflow‑builders loved the idea of using Showboat as “gates” that block tasks until a demo passes.

Add in some side‑drama over Simon shipping Go binaries (cue the Python‑vs‑Go whispers), and you’ve got classic HN: useful idea, instant debate, and a few “showboating bots” jokes for dessert.

Key Points

  • Simon Willison released Showboat and Rodney to help coding agents demonstrate and prove their work.
  • Showboat is a CLI Go binary (optionally wrapped in Python) that generates Markdown documents showing software capabilities.
  • The tools aim to reduce manual QA and minimize opportunities for agents to misrepresent their progress.
  • Willison references StrongDM’s software factory model, which uses QA agent swarms and avoids human code review, as a costly alternative.
  • An example artifact shows commands using curl and jq, captured outputs, and embedded images previewed in VS Code.

Hottest takes

"could be one shot by either Claude or Codex" — saberience
"Closing the loop on a web interface isn't as simple as CLI-only tools" — eliben
"It doesn’t even generate the correct syntax block" — toastal
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.