voxspek
← Back to voxspek
Legal

Privacy Policy

Last updated: 2026-04-10

What we collect

The short version

Your API keys never leave your machine. They live in chrome.storage.local on your own browser and are transmitted directly to Anthropic, OpenAI, and GitHub via their official APIs. We do not run any server that receives your keys. You can verify this yourself by inspecting the open-source code or using Chrome's developer tools to watch network traffic.

When you press the hotkey and submit a recording, voxspek captures:

  1. Your voice recording — an audio file of what you spoke
  2. The URL of the current tab — the page you were on
  3. The page title
  4. A DOM snippet — up to 4 KB of HTML around the focused element
  5. Recent console log entries — up to 50 messages from the browser console
  6. Your API keys (stored locally, never transmitted beyond the services you use)

What we do with it

The captured data is sent to:

  1. Your own Vercel serverless function (the API base you configured) — the data transits this endpoint to be processed.
  2. OpenAI Whisper — your audio recording is transcribed here.
  3. Anthropic Claude — the transcript plus context is sent here to produce a structured issue.
  4. GitHub — the resulting issue is posted here.

voxspek does NOT store any of this data on its own servers. We do not have any servers. The only places your data touches are the four services listed above, all of which you control via your own API keys.

What we do NOT collect

  • No analytics
  • No telemetry
  • No tracking
  • No user accounts
  • No third-party cookies
  • No data retention beyond what OpenAI, Anthropic, and GitHub store per their own policies

Your responsibility

Because voxspek captures whatever is on your screen when you press the hotkey, you should:

  • Avoid using it on pages containing sensitive personal information (financial data, medical records, etc.) unless you trust the downstream services with that data
  • Configure a "paused domains" list if needed
  • Review OpenAI's, Anthropic's, and GitHub's privacy policies — they process your data under their own terms

Data retention at third parties

  • OpenAI Whisper: per OpenAI's API policy, audio is not used for training and is retained per their data retention policy
  • Anthropic Claude: per Anthropic's API policy, inputs are not used for training
  • GitHub: issues you create are retained per GitHub's terms — they are GitHub issues, treat them accordingly

Contact

This is a solo-developer tool built by Etienne Bernard. Questions or concerns: etienne@nora-strategy.com