agentleFS
Omodaka9375Omodaka9375/browserr/docs/llms.txt

browserr

llms.txt58 starsChanged yesterdayApache-2.0View on GitHub

A browser engine for AI agents. One Rust binary built on Servo: no CDP, no Chromium, and a small protocol you can audit in an afternoon. The MCP server exposes the engine to any MCP client; the engine fails closed, and human handoff (passwords, captchas, payments) happens in an optional native window, never through the agent. Browserr drives real pages - navigation, reading, forms, clicks, downloads, interception - with an evidence trail: every action is logged and every screenshot is…

The file

Files
llms.txt
# Browserr

> A browser engine for AI agents. One Rust binary built on Servo: no CDP,
> no Chromium, and a small protocol you can audit in an afternoon. The MCP
> server exposes the engine to any MCP client; the engine fails closed, and
> human handoff (passwords, captchas, payments) happens in an optional
> native window, never through the agent.

Browserr drives real pages - navigation, reading, forms, clicks, downloads,
interception - with an evidence trail: every action is logged and every
screenshot is written to disk before the result is returned. Version 0.7.x.

## Docs

- [Install](https://github.com/Omodaka9375/browserr#install): Prebuilt binaries for Windows x64, Linux x64/ARM64, macOS ARM64.
- [Protocol](https://github.com/Omodaka9375/browserr#protocol): The 22-command JSON-lines-over-TCP protocol, every command and event documented.
- [MCP server](https://github.com/Omodaka9375/browserr#mcp-server): The 42 tools; every tool accepts a named engine session.
- [Profiles](https://github.com/Omodaka9375/browserr#profiles): Persistent profiles keep logins and storage across sessions.
- [Security notes](https://github.com/Omodaka9375/browserr#security-notes): The trust model: evidence directory, action log, confirm gates for payments, fail-closed interception.
- [Known limitations](https://github.com/Omodaka9375/browserr#known-limitations): What Browserr does not do yet.
- [Tests](https://github.com/Omodaka9375/browserr#tests): The 42-test suite and how to run it.

## Optional

- [Terminal client](https://github.com/Omodaka9375/browserr/tree/main/term): Interactive TUI rendering native Sixel/Kitty frames from the same engine.
- [Driver SDK](https://github.com/Omodaka9375/browserr/blob/main/tools/driver.py): The Python client the MCP server and the terminal client are built on.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.