Setting up
Three short steps, about five minutes in total. There is no account, no API key and no cloud service.
- Install — clone, build, load the extension
- Pair — put
browsentic-mcpon yourPATHand connect your browser to it - First run — a tour of the side panel, and your first instruction
Using it
Features — one page per capability:
| Conversations | Voice, text, one conversation per tab, history |
| Instant commands | The things that run in milliseconds without an agent |
| Page actions | What it can actually do to a page |
| Screenshots | Capturing the viewport, the full page, or one element |
| Theming | Dark mode on any site, and a real contrast audit |
| Captchas | What it will and will not do at a "verify you are human" block |
| Monitoring | Watching a long job in the background |
| Site maps | Teaching it a site once |
| Recordings | Showing it a task once, repeating it later |
| Files | Attaching a file and uploading it to a page |
| Skills | How instructions get routed, and how to write your own |
Configuring
| Choosing an agent | Claude Code, Codex or Antigravity — switching, and what each needs |
| MCP clients | Registering Browsentic with Claude Code, Cursor, Zed, Codex, Gemini CLI |
| Configuration | Every key in ~/.browsentic/config.json |
| Approvals | The gate — what asks first, and how to tune it |
When something is wrong
| Limits | The honest boundaries. Read this before relying on it for anything |
| Troubleshooting | Symptom, cause, fix |
| Maintenance | Updating, and removing it cleanly |
Want to know how it works underneath? Internals.