</>MCP Agents Market
MCP Server

workflow-use

by browser-use4.2kPythonUpdated 2026-08-28

⚙️ Create and run workflows (RPA 2.0)

Workflow Use is an MCP server that enables developers to create and execute deterministic browser automation workflows with self-healing capabilities. It allows you to record browser interactions once and replay them indefinitely, with automatic fallback to Browser Use AI agent when steps fail. Workflows can be generated from natural language descriptions, stored in a database, and reused with different inputs without requiring AI calls. The server integrates with Browser Use and supports both local and cloud browser execution.

Key Features

Record browser interactions once and convert them into reusable deterministic workflows
Generate workflows automatically from natural language task descriptions using Browser Use
Self-healing workflows that fall back to AI agent execution when deterministic steps fail
Workflow storage and retrieval system with metadata indexing and search
Run workflows with no AI calls using semantic parameter mapping
Visual GUI for managing, visualizing, and executing workflows as interactive graphs
Cloud browser support via Browser Use Cloud integration
Programmatic API and CLI for workflow generation, execution, and management

Use Cases

  • 01Automate repetitive form-filling tasks on websites with variable data
  • 02Create reusable web scraping workflows that extract structured data reliably
  • 03Build RPA 2.0 workflows for enterprise automation with self-healing capabilities
  • 04Generate workflow templates from one-time AI browser sessions for repeated execution
  • 05Integrate deterministic browser automation into AI agent toolchains via MCP
  • 06Monitor and execute complex multi-step browser workflows through a visual interface

Related MCP Servers

View more

workflow-use — FAQ

What is Workflow Use MCP server?+

Workflow Use is an MCP server for creating and running deterministic browser automation workflows that self-heal using AI. It allows you to record browser interactions once, then replay them reliably with different inputs, falling back to the Browser Use agent when steps fail.

How do I install Workflow Use?+

Clone the repository, build the browser extension with npm, then set up the Python environment using uv sync and install Playwright's Chromium browser. You'll need to add your OpenAI API key to a .env file before running workflows.

What are the prerequisites for using Workflow Use?+

You need Node.js and npm for the extension, Python with uv package manager, Playwright with Chromium installed, and an OpenAI API key. Optionally, a Browser Use Cloud API key enables cloud browser execution with free credits for new signups.

Which AI clients work with Workflow Use?+

Workflow Use is designed to work as an MCP server building block and integrates with Browser Use. The roadmap indicates it will expose workflows as MCP tools for broader AI client compatibility.

Is Workflow Use free to use?+

The software itself is open source and free. You'll incur costs for OpenAI API usage when generating workflows or using self-healing features. Browser Use Cloud offers $10 free credits for new OAuth signups or $1 via email.

Can I run workflows without AI after recording them?+

Yes, once recorded or generated, workflows can execute deterministically without any LLM calls using semantic parameter mapping. AI is only used during initial workflow generation or when self-healing is triggered by step failures.

How do I install workflow-use?+

Open the source repository on GitHub and follow its README. workflow-use is a mcp server — MCP Agents Market links you directly to the official repo.

Is workflow-use free?+

workflow-use is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches