Free tool · no signup
Eval a prompt in your browser, with no signup
On this page
The short answer
- Paste a system prompt and run a review in your browser. No account.
- Offline heuristic mode is the default. It is clearly labelled and does not call a model.
- Live mode is optional and uses your own API key, sent only to the provider you choose. Litmiq never receives it.
Prompt checker
The two modes
Offline heuristic mode looks for empty or tiny prompts, leftover placeholders, a stated role, limits, an output shape, a source of truth, and an example when the format is strict. Each check is pass, warn or fail, with a reason. The 0–10 figure is only a summary of those checks.
Live mode asks a model you pay for to review the same prompt, then still shows the offline checks. It is one model’s opinion, not a Litmiq benchmark and not a substitute for a suite of expected-behavior cases.
Limits
This tool does not run your test cases, does not store history, and does not replace LLM-as-a-judge evals in the app. For injection and PII refusals, use the red-team guide and the safety template.
FAQ
Is the offline score an LLM judge?
No. Offline mode is a labelled heuristic: length, placeholders, role, limits, output shape, source of truth and examples. It does not call a model and it is not a quality guarantee.
Where does my API key go in live mode?
Only to the provider you choose, from this browser. Litmiq has no server that receives the key or the prompt. Some providers block browser requests; if that happens, use offline mode or open the prompt in the Litmiq app.
Do I have to create an account?
No. The checker runs on this page. A key is optional and only needed for live mode.
Can I run a full suite?
Yes. Use a template from the prompt-templates library, or open the Litmiq app and add your own cases.
Run a full suite
Heuristics are a start. A template gives you expected-behavior cases.
- Free, no signup
- Runs in your browser
- Your OpenAI key, sent only to OpenAI