Copy as Context per V-Shkrobatskyi

Copy the current page as compact, structured context for an AI conversation or local notes. Copy as Context processes page content locally and exports it as Semantic Text or Markdown. No account, API key, or paid service is required.

Watch the demo video on YouTube.

Features:
- Four compression levels: Without, Detailed, Compact, and Maximum.
- Semantic page structure, including headings, links, lists, tables, and supported form controls.
- Iframe capture with embedded document positions and boundaries preserved in exports.
- Credential redaction enabled by default.
- Saved compression, format, and privacy preferences.
- Character counts, approximate token counts, and context reduction.

Firefox Desktop supports Copy and Save as a local UTF-8 text or Markdown file. Save requests optional download permission and opens a separate extension tab that closes after completion.

Firefox for Android supports Copy. File saving is unavailable on Android.

There is no LLM integration, capture backend, analytics, or advertising. The extension does not upload captured content to the developer. You choose where to paste or share the result.

Capture uses HTML and ARIA semantics and includes available same-origin frames by default. Enable "Include frames from other sites" to request optional access to all HTTP/HTTPS websites. This permission remains granted until revoked; disabling the option removes site access. Capture runs only when you click Copy or Save. Hidden frames are skipped, and unavailable embedded documents are reported.

Capture does not perform OCR or include closed shadow roots, canvas pixels, or content that has not loaded. Browser-internal and protected pages may deny access.

Password input values are always omitted. Credential redaction is heuristic and may miss sensitive information. Review exports before sharing them. Token counts are estimates and depend on the model.