ChatGPT Browser Chat Telemetry in Beacon

What is ChatGPT?

ChatGPT usage often sits next to coding work even when no local agent process is running. Beacon treats that browser activity as a supported surface with clear scope and local retention controls.

Collection path: Managed Chrome MV3 extension over local OTLP

Authoritative date: August 24, 2026 from Browser extension in the Beacon CLI changelog.

How Beacon reads ChatGPT

The optional Chrome extension reads supported chat streams on chatgpt.com and chat.openai.com, then sends normalized OTLP GenAI logs to Beacon local collector. Events are attributed to chatgpt_web, not a CLI harness.

The result is normalized into Beacon's endpoint event schema so security teams can investigate activity by harness, repository, session, user, model, command, file, tool, MCP server, or approval context where the runtime exposes those fields.

Telemetry coverage

• Prompt telemetry from supported ChatGPT browser chat streams
• Assistant response telemetry where retained by policy
• Tool call telemetry where exposed by the browser chat stream
• Local runtime.jsonl output and existing customer-managed forwarding paths

What you can do now

• Bring ChatGPT browser usage into the same AI activity record as endpoint agents.
• Query browser chat separately from CLI and IDE agents with the chatgpt_web harness.
• Use metadata-only or redacted retention when work and personal profiles overlap.

Get started

1. Update to a Beacon release that includes browser extension support.
2. Build the extension from browser-extension/ and load it unpacked in Chrome.
3. Set retention in the extension options before enabling collection on shared profiles.

Source references

• Runtime docs
• Source reference