Free tools for building AI apps

Switch to a better, cheaper model without your users noticing, and the other questions builders ask once their AI feature meets real users. Each tool runs in your browser and needs no account.

• How often does your AI judge agree with a person?

  Check how often your AI judge agrees with a person, and whether to trust it.

  Judge agreement checker. Builds a test set from real requests, with an AI judge to score it.

• Write a judge that checks your AI's results

  A pass/fail judge prompt for any AI step, from support replies to invoice fields.

  Judge prompt builder. Builds a test set from real requests, with an AI judge to score it.

• Switch to a better, cheaper model without your users noticing

  Run your prompt on two models side by side and compare what worked and what it really costs.

  Model switch test. Tries a new prompt or model on some of your users before everyone gets it.

• What does each AI result really cost?

  What each AI result really costs, what you waste a month, and what to charge in credits.

  Real cost per result calculator. Records what every AI step costs, and your margin on each result.

• How many results before you know version B works first time more often?

  How many results you need before a prompt or model change is proven.

  Prompt A/B test sample size. Tries a new prompt or model on some of your users before everyone gets it.

• Is your vibe-coded AI app ready for paying users?

  The AI-specific checks that generic vibe-coding checklists skip.

  AI launch checklist. Records which AI results worked first time and which needed a try again.

• Which AI models are being shut down, and when?

  Announced API shutdown dates and replacements, with sources.

  Model shutdown calendar. Tries a new prompt or model on some of your users before everyone gets it.

Also free: AGENTS.md sections for 8 kinds of AI app and a price page for every AI model, re-checked daily.