Browser Chat Telemetry for Claude.ai and ChatGPT

Beacon now treats browser AI chat as a supported surface, bringing Claude.ai and ChatGPT conversations into the same local pipeline as endpoint agent activity.

August 24, 2026

What is browser chat telemetry?

Not every AI-assisted workflow starts in a terminal or IDE. Developers also use Claude.ai and ChatGPT in the browser to reason about code, paste snippets, inspect errors, and draft changes before moving back into their repositories.

Beacon's browser extension brings those conversations into the endpoint record, so browser chat no longer sits outside the system of record for AI activity.

How Beacon reads browser chat

Beacon uses an optional Chrome MV3 extension. The extension relays supported chat streams into the collector already listening on 127.0.0.1:4318, where they are written to runtime.jsonl beside the rest of the agent activity.

The extension is narrow by construction. It reads chat streams on claude.ai, chatgpt.com, and chat.openai.com. It has no access to other tabs, browsing history, or page content elsewhere, and it never writes files.

Activity is attributed to the claude_web and chatgpt_web harnesses, keeping browser chat distinguishable from CLI and IDE agents.

What you can do now

With this update, teams can:

Capture Claude.ai and ChatGPT conversations in the same local runtime log used by endpoint agents.
Forward browser chat telemetry through the same configured shippers as the rest of Beacon's endpoint data.
Configure retention as full, redacted, or metadata-only before enabling the extension on shared browser profiles.

Get started

Update to the current Beacon endpoint release.
Build the extension from browser-extension/ and load it unpacked in Chrome.
Choose the retention setting before enabling it on a machine where personal and work conversations share the same browser profile.