Privacy Policy
Last updated: March 1, 2026
Overview
Kesit is an agentic code-search tool that runs entirely on your Mac. We are committed to protecting your privacy. This policy explains how Kesit handles your data.
Data Collection
Kesit does not collect, transmit, or store your screen content or extracted code on any external server. All code extraction and AI processing runs locally on your device.
Screen Recording
Kesit uses macOS screen recording permission to capture code regions you select. Captures are initiated only by your explicit action — pressing the global hotkey (⌥ Space), clicking the Capture button, or pasting an image from clipboard. No background, continuous, or automatic screen recording occurs.
The captured image is processed entirely on your device using Apple Vision OCR and, on supported systems, Apple Intelligence via FoundationModels. No screenshot data, extracted code, or derived information is ever transmitted off your device.
On-Device AI Processing
Kesit uses Apple's native on-device frameworks for all AI functionality:
- Apple Vision — Recognises text from captured screen regions using VNRecognizeTextRequest. All OCR processing runs locally via the Vision framework.
- Apple FoundationModels (macOS 26+) — Cleans OCR artifacts, fixes indentation, and powers agentic Skills (translate, document, explain) using Apple Intelligence. This runs entirely on your Mac's Neural Engine. On systems that do not support Apple Intelligence, deterministic code extraction is used instead.
Kesit does not use any third-party AI services. No data is sent to OpenAI, Anthropic, Google, or any other external AI provider. All inference happens on-device.
Data Storage
Kesit stores the following data locally on your Mac:
- Snap history — Extracted code snippets, language tags, source window context, and timestamps are persisted in a local SwiftData database in your app container. You can browse, search, and delete snaps at any time.
- User corrections — If you edit OCR results, the corrections are saved locally for few-shot learning context.
- Preferences — Window size, capture settings, and app configuration stored via standard macOS mechanisms.
No snap data, code extractions, or user corrections are stored on any server, cloud service, or external location.
Analytics & Crash Reporting
Kesit includes two optional, privacy-respecting services:
- Sentry — Crash reporting to help us identify and fix bugs. Sentry receives crash stack traces and basic device metadata (OS version, app version). It does not receive any screen content, extracted code, or snap data.
- TelemetryDeck — Anonymous usage analytics to help us understand which features are used. TelemetryDeck collects non-identifying event counts (e.g., "capture triggered", "skill used") with no personal identifiers, no IP address retention, and no screen content.
Both services are designed to be privacy-first. Neither receives your code, screenshots, or any personally identifiable information. You can opt out of analytics in Kesit's settings.
Third-Party Services
Apart from the optional crash reporting (Sentry) and usage analytics (TelemetryDeck) described above, Kesit does not integrate with, send data to, or receive data from any third-party services. There are no advertising frameworks or tracking libraries embedded in the app.
Children's Privacy
Kesit does not collect any personal information from anyone, including children under 13.
Changes to This Policy
If we update this privacy policy, we will post the revised version on this page with an updated date.
Contact
If you have any questions about this privacy policy, you can reach us at hi@alp.me.