An open-source AI browser for autonomous agents with durable state, MCP control, and human supervision.
vessel-browser is worth checking the docs before setup with trust notes worth reviewing. Check agent compatibility and use-case fit before adding it to your workflow.
Repository setup guidance
gh repo view unmodeled-tyler/vessel-browser --webOpen the official repository or website.
Check the README for package manager, auth, and platform requirements.
Try it in a small test task inside your agent workflow.
Vessel Browser is a web browser built for AI agents, not humans. It keeps browser sessions alive, allows agents to control it via MCP, and lets humans watch and intervene when needed. It works with agent frameworks like Hermes Agent and OpenClaw.
Vessel Browser is an open-source, Chromium-based browser designed from the ground up for autonomous AI agents. Unlike traditional browser automation tools that are headless or stateless, Vessel provides a persistent, stateful browser environment that agents can control via the Model Context Protocol (MCP). It features named sessions, pinned tabs, editable bookmarks, annotated checkpoints, action undo, and a visible supervisory UI for humans. Built for long-running agent workflows, Vessel keeps humans in the loop with approval requests, runtime controls, and full visibility into agent actions. It supports Linux, macOS, and Windows, and can be installed via AppImage, npm, or built from source. Vessel is compatible with agent harnesses such as Hermes Agent, OpenClaw, and any MCP client. The project is in active development and currently makes no security assurances.
Looks usable, but maintenance, license, or security notes deserve a closer look.
Last commit was about 51 days ago.
88 GitHub stars indicate community interest.
3 open issues signal maintenance load.
MIT license detected.
Run long-duration web automation tasks with persistent browser state and human oversight.
Enable AI agents to perform complex multi-step web interactions like form filling, data extraction, and navigation.
Provide a visible, auditable browser interface for agent-driven research or testing workflows.
Integrate with MCP-compatible agent frameworks to give them a real browser environment.
No security assurances are provided; use with caution in production environments.
The browser runs with full system access; agents could potentially execute harmful actions if not properly supervised.
88
Stars
11
Forks
3
Issues
MIT
License
Extract any website's complete design system with one command.
Argos is an open source visual testing platform that detects unintended UI changes to help teams maintain quality.
Generate design system documentation for UI components directly from your AI agent, rendering into Figma or portable .md files.
An open-source AI browser for autonomous agents with durable state, MCP control, and human supervision.
AI/ML engineers building autonomous agent systems that need a durable browser environment., Developers creating MCP servers or clients that require a full-featured browser for agent interactions., Researchers and QA engineers who need to observe and supervise agent-driven web automation.
The setup section provides repository-level starting guidance, not a guarantee of an independently verified installation. Check the official README and release notes.
No security assurances are provided; use with caution in production environments. The browser runs with full system access; agents could potentially execute harmful actions if not properly supervised.
Similar or complementary options to evaluate include design-extract, Argos, uSpec.
Jun 7, 2026
2 security/trust notes recorded.
Setup difficulty is 3/5.