Fieldbook to AIThe plain-English guide to which AI to use

Time-savers

Get a first check on code changes (for developers)

Every time someone proposes a code change, a model reads it and leaves a short comment on likely problems before a person looks.

Last updated .

Costs here use the prices AI companies charge software developers, so they show relative cost, not your monthly bill. See what ChatGPT, Claude and Gemini charge you.

Starter recipe, not yet tested end to end.

What it does for you

Reads a proposed code change and lists real problems only: bugs, security issues and missing tests, each with where and how to fix it.

Time saved

Catches obvious mistakes early so human reviewers spend time on the important parts. We have not measured the time saved yet.

What it costs to run

Model to start withCost for 100 code changes
Claude Sonnet 5.5$4.00
Claude Opus 5.5$8.00
Muse Spark Open$2.30

Code review rewards the strongest model you can afford; the downloadable option suits teams that must keep code in-house. Estimate assumes about 15,000 tokens in and 1,000 tokens out per item (a token is about three-quarters of a word), at list prices. Real usage will differ.

How to build it

You will need: GitHub Actions; Your code repository

  1. Start when a code change is opened or updated.
  2. Collect the change and the files around it.
  3. Send it to the AI model with the instructions below.
  4. Post the result as one clearly labeled automatic comment.

Instructions to give the AI

You are a careful code reviewer. Review this change. List only real problems: bugs, security issues, missing tests. For each, give the file, line and a one-sentence fix. If it looks fine, say so. Change:
{{diff}}