Privacy Policy for PromptLens

Effective Date: July 26, 2026

PromptLens is committed to respecting user privacy. This Privacy Policy explains how PromptLens processes data, why specific browser permissions are required, how local storage is used, and the handling of user-initiated third-party AI context routing.


1. Local Processing Architecture

PromptLens is designed around a local-first processing model:


2. Data Handling, Storage & Retention


3. Browser Tab Information & AI Destination Routing

PromptLens inspects open browser tabs locally to identify active AI assistant destinations:


4. Extension Permissions & Technical Justifications

PromptLens requests only the permissions necessary to deliver its core single-purpose features:

Permission User-Facing Technical Justification
activeTab Allows PromptLens to capture the visible page region when you explicitly trigger a screenshot or OCR selection.
tabs Allows PromptLens to inspect open tab URLs locally to discover active AI destinations (ChatGPT, Claude, Gemini, Grok, Perplexity) for prompt routing.
scripting Allows PromptLens to render visual selection overlays and floating recording controls on the active page.
offscreen Allows PromptLens to run WebAssembly OCR extraction and tab recording streams in an isolated background document.
storage Allows PromptLens to persist local extension state and active AI destination tab options using chrome.storage.local, while session capture history is managed separately in memory via chrome.storage.session.
tabCapture Allows PromptLens to capture browser tab video streams when you explicitly start a screen recording workflow.
clipboardWrite Allows PromptLens to copy extracted OCR text or screenshots to your system clipboard upon your explicit action.
<all_urls> Required to display the interactive visual selection overlay across web pages when you press a hotkey or trigger capture.

5. Third-Party Data Sharing & Limited Use Compliance


6. Open-Source Verification & Contact

PromptLens is an open-source project. Users and security researchers may inspect the complete source code, permission definitions, and local processing implementation on GitHub: