Privacy-first Effective April 16, 2026

Privacy Policy

Bonsai Capture is built to keep your AI conversations local. The extension stores captures on your device, and this landing page uses privacy-preserving Cloudflare Web Analytics plus the information you choose to submit through the waitlist form.

At a glance

  • Extension captures are stored locally in Chrome storage.
  • The site uses Cloudflare Web Analytics for aggregate traffic measurement.
  • No advertising pixels, remarketing tags, or data sales are used on the site.
  • Waitlist form submissions are sent only if you choose to submit them.

Website

What the landing page collects

The Bonsai Capture landing page is a static marketing site. It uses Cloudflare Web Analytics to understand aggregate traffic and referral trends. It does not use advertising pixels or remarketing tags. If you submit the waitlist/contact form, the name, email, and message you provide are transmitted to Formspree so the request can be delivered.

If you do not submit the form, the site does not collect personal information.

Extension

What the extension accesses

The extension reads the content of supported AI chat pages so it can capture conversation text, structure, code blocks, artifacts, inline images, and generated media that you intentionally open in those tabs.

Captures are stored in chrome.storage.local on your device and are not uploaded to any remote service by default.

Permissions

Why the extension needs elevated permissions

activeTab, scripting, sidePanel

Used to open the side panel and inject capture logic only when you actively use the extension on supported pages.

storage, unlimitedStorage

Stores captures locally, including conversation records and large embedded media data URLs.

declarativeNetRequest

Creates temporary per-request header rules for provider-hosted media fetches that require Origin and Referer headers. Rules are removed immediately after each fetch.

clipboardRead

Fallback for capturing Claude artifact source when the native copy flow is the only reliable extraction path. It is used only on supported capture flows.

Host access

Why specific domains are listed

Content scripts run only on ChatGPT, Claude, Gemini, and Grok pages so the extension can read the conversation DOM. Additional host permissions are limited to provider CDNs that host generated media or sandboxed research frames that the capture flow must access.

Retention

How long data is kept

The extension keeps a rolling window of the most recent 50 captures. Older captures are dropped automatically. You can also clear the data by removing extension storage or uninstalling the extension.

The landing page does not retain waitlist submissions itself; Formspree handles form delivery for submitted entries.

Contact

Questions or concerns

For questions about this policy or the extension, open an issue in the project repository or use the site’s contact/waitlist form if you want a direct reply.