Browser Use Web UI

Gradio web UI for running browser-use AI agents in a real or persistent Chrome session.

Free· Free, open-source; bring your own LLM API keysAgentsMulti-model

Best for

Pick Browser Use Web UI if you want a quick GUI to prototype LLM-driven browser agents against your own Chrome profile and chosen model.

Skip if

Skip it if you need a hosted, managed agent platform with SLAs, audit logs, or non-technical operator tooling.

Browser Use Web UI is an open-source Gradio frontend that wraps the popular browser-use Python library, letting you point an LLM at a real browser and have it drive web pages on your behalf. It exposes most of browser-use's agent functionality through a friendly interface, supports persistent browser sessions so the agent can keep state between tasks, and lets you attach your own logged-in Chrome profile to skip re-authentication on sites you already use.

It is aimed at developers and tinkerers who want to experiment with browser-controlling agents without writing glue code. The project is fully self-hosted (local Python install or Docker with a VNC viewer at port 6080), and it is LLM-agnostic: OpenAI, Anthropic, Google, Azure OpenAI, DeepSeek, and Ollama are all wired in, so you bring your own API keys (or run a local model) and pay only for inference. There is no SaaS tier and no usage cost beyond the underlying LLM.

The UI also supports HD screen recording of agent runs, which is genuinely useful for debugging brittle multi-step web automations. Caveats: it inherits browser-use's fragility on heavily client-rendered or anti-bot-protected sites, and the Gradio interface is functional rather than polished. Best treated as a power-user playground on top of the browser-use ecosystem.

Editor's take

A pragmatic companion to the browser-use library: same capabilities, but with a Gradio UI and Docker+VNC story that make it the fastest way to actually see a browser agent run. Treat it as a developer playground rather than a shippable product, and pair it with a strong LLM for usable results.

— The AI Tool Bible editorial team

Pros

✅ Runs a real browser under LLM control, with optional persistent profile and logged-in cookies
✅ LLM-agnostic: OpenAI, Anthropic, Gemini, Azure, DeepSeek, Ollama all supported
✅ Docker setup with built-in VNC viewer for watching the agent work
✅ Fully open-source and self-hosted; no per-task fees beyond the LLM