Open-source browser automation that uses LLMs and computer vision to run web workflows on any site, extending Playwright with AI commands.
browser-use/workflow-use
Records browser interactions once and replays them as deterministic workflows, falling back to Browser Use AI when a step fails.
Open-source alternative to:UiPath
- Stars
- 4,247
- Forks
- 356
- Open issues
- 72
- Last commit
- yesterday
What is it?
Workflow Use turns a browser session into a reusable, deterministic workflow. You perform a task once in a Chrome extension recorder, and the tool generates a JSON workflow with variables extracted from form inputs. Running the workflow replays the steps directly without calling an LLM, which keeps execution fast and predictable.
When a step fails, the workflow can fall back to Browser Use, an AI browser agent, to complete it. A generation mode also lets you describe a task in natural language, runs Browser Use once, and converts that execution history into a semantic workflow stored in a local database for later reuse with different inputs.
The project includes a CLI for recording, generating, listing, and running workflows, a Python API, Browser-Use Cloud support, and a web GUI built with FastAPI and a React-style frontend for viewing and executing workflows. It is in early development, and the README advises against production use.
Key features
- 01Records browser interactions via a Chrome extension and converts them into JSON workflows.
- 02Automatically extracts form inputs as workflow variables for reuse with different values.
- 03Replays deterministic steps without LLM calls, with Browser Use fallback when a step fails.
- 04Generates workflows from natural language task descriptions and stores them in a database.
- 05Runs workflows locally or on Browser-Use Cloud browsers.
- 06Provides a CLI, a Python API, and a web GUI for managing workflows.
Use cases
- Repeating routine web form submissions with different data each time.
- Automating multi-step web tasks that should behave the same way every run.
- Prototyping browser automation by recording a task instead of writing selectors by hand.
- Running recorded browser workflows as tools for other agents or scripts.
Quick start
- 1
Clone the repository
git clone https://github.com/browser-use/workflow-use - 2
Sync the workflow environment dependencies with uv
uv sync - 3
Install the Chromium browser for Playwright
playwright install chromium - 4
Run the example workflow
python cli.py run-workflow examples/example.workflow.json
Things to know
- The project is in very early development and not recommended for production use, with breaking changes expected.
- Licensed under AGPL-3.0, a copyleft license that requires sharing modifications when distributed or served over a network.
- The AI fallback relies on an OpenAI API key set in the .env file.
In collections
Topics
FAQ
- What is workflow-use?
- Records browser interactions once and replays them as deterministic workflows, falling back to Browser Use AI when a step fails. Workflow Use turns a browser session into a reusable, deterministic workflow. You perform a task once in a Chrome extension recorder, and the tool generates a JSON workflow with variables extracted from form inputs. Running the workflow replays the steps directly without calling an LLM, which keeps execution fast and predictable.
- Is workflow-use free to use? What license does it have?
- workflow-use is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
- Is workflow-use still maintained?
- Actively maintained. The last commit to workflow-use was on Nov 6, 2026.
- What can I use workflow-use for?
- Repeating routine web form submissions with different data each time. Automating multi-step web tasks that should behave the same way every run. Prototyping browser automation by recording a task instead of writing selectors by hand. Running recorded browser workflows as tools for other agents or scripts.
- How do I install workflow-use?
- Clone the repository: git clone https://github.com/browser-use/workflow-use Sync the workflow environment dependencies with uv: uv sync Install the Chromium browser for Playwright: playwright install chromium Run the example workflow: python cli.py run-workflow examples/example.workflow.json
- Where can I get workflow-use?
- The source code and installation instructions are on GitHub: https://github.com/browser-use/workflow-use
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source enterprise RPA desktop app with a visual low-code designer for automating Windows desktop software and web pages, callable from AI agents.
Open-source execution layer that records a task once and compiles it into a local program agents or people can rerun with no model calls, verifying each save.
Python and Robot Framework libraries for Robotic Process Automation, covering browsers, Excel, email, PDF, cloud APIs and desktop apps.
Browser and desktop automation extension with a macro recorder, JavaScript API, OCR, image recognition, and an MCP server for AI assistants.
Free, open-source RPA tool that automates websites, desktop apps and command-line tasks with simple TagUI flow scripts.