From €400, delivered in 2–4 days. A written review of safety gaps, cost, key handling and privacy, with fixes.

I review an LLM feature you already have, or are about to ship, and give you a written report with concrete fixes. What I check - Safety: what happens with harmful, confusing or off-topic messages, and whether the model's reply is checked before the user sees it. - Keys and secrets: how provider keys are stored, passed and kept out of logs, the repository and error messages. - Privacy: whether user text reaches logs, analytics or third parties, and what is stored on disk. - Cost and reliability: retries, limits, long inputs, and what a failed or empty model answer does to the user. - Tests: whether the feature can be tested without a paid model.

What you get: a written report ranked by severity, with a fix for each finding and a short call or message thread to go through it. What I need from you: read access to the relevant code, or a walkthrough of the feature, and a note on which models and providers you use. Delivery: 2-4 days, depending on the size of the feature. Fixed price. Good to know: this is a review, not a guarantee that nothing can go wrong. I don't need your live keys and won't ask for them. Please don't send them. I build and test this kind of feature myself: my apps Socrates and Moly use the user's own key, a safety check on every message and reply, encrypted local storage and no user text in logs.