LLM Bench Privacy Policy

Effective date: 2026-06-02

LLM Bench is an iOS developer tool for repeatable on-device GPU LLM benchmarking. This policy explains what data the app keeps local, what limited analytics may be sent, and how to contact support.

Local Benchmark Data

LLM Bench is designed around local benchmarking.

The app does not collect your prompts, model responses, benchmark output, model files, imported model paths, or saved run history through usage analytics. Benchmark data is stored on your device unless you choose to export it, share it, or include it in feedback.

Model Files

Model files you download or import are stored locally on your device. LLM Bench may download model artifacts from URLs you provide or from supported model sources you choose in the app.

LLM Bench does not upload your model files.

Usage Analytics

LLM Bench may collect limited usage analytics to understand app health and improve the product. Examples include high-level events such as onboarding progress, paywall views, purchase results, benchmark activity, model download activity, custom eval usage, and CSV export usage.

Usage analytics does not include prompts, model responses, model files, benchmark outputs, or saved run history.

Usage analytics can be turned off in Settings.

Screen Capture And Session Replay

LLM Bench does not enable PostHog screen capture, session replay, automatic screen-view capture, element autocapture, or rage-click autocapture.

Purchases

LLM Bench uses Apple StoreKit for subscriptions and lifetime purchase options. Apple processes purchases and provides purchase status to the app.

Feedback And Support

If you contact support, you choose what information to send. Feedback may include device, memory, battery, runtime, app version, and build diagnostics when you include them.

Contact: dansedano.dev@gmail.com

Changes

This policy may be updated as the app changes. The latest version should be published at the privacy policy URL listed in App Store Connect.