Skip to content

Use an AI agent without becoming its QA department, even when you are the only person who checks the result

Excellent runs beside the agent, checks the work against evidence you can read, and hands back a receipt instead of a confident summary.

curl -fsSL https://excellent.so/install.sh | sh

What you recognise

  1. 01

    The agent produces a plausible patch faster than you can prove whether it works, so you read the whole diff anyway.

  2. 02

    You rerun the same commands by hand after every attempt and still miss the case you did not think to check.

  3. 03

    When an approach finally works, nothing records why. The next task starts from scratch.

What changes

The claim and the check are separate

The agent's "done" is recorded as a claim. What ran, what changed, and what passed is recorded separately, so you read a result rather than a summary.

Failures come back with the failure in them

Returned work names the unmet check and the observed output, so the next attempt starts from evidence instead of a re-prompt.

It stays on your machine

Attempts, artifacts and receipts are written locally. There is no account to create and nothing to upload before you can see a result.

See for yourself

Where Excellent stops

  • Runs on macOS, with Node.js 22.3 or newer.
  • There is no desktop app to download. Excellent is a command-line tool today.
  • Browser, CI and deployment evidence are experimental. Git and test evidence are the solid ground.

Someone else does this work at your company?

Run one real task and see what held up

curl -fsSL https://excellent.so/install.sh | sh