Stepflow

Open protocol and runtime for building GenAI workflows
Explore the docs »

View Demo · Report Bug · Request Feature

Table of Contents

1. About The Project
2. Getting Started
   • Prerequisites
   • Installation
3. Usage
4. Roadmap
5. Contributing
6. License
7. Contact

About The Project

Stepflow is an open protocol and runtime for building, executing, and scaling GenAI workflows across local and cloud environments. Its modular architecture ensures secure, isolated execution of components—whether running locally or deployed to production. With durability, fault-tolerance, and an open specification, Stepflow empowers anyone to create, share, and run AI workflows across platforms and tools.

Key Features

• ⚙️ Reliable, Scalable Workflow Execution Run workflows locally with confidence they'll scale. Stepflow provides built-in durability and fault tolerance—ready for seamless transition to production-scale deployments.
• 🔐 Secure, Isolated Components Each workflow step runs in a sandboxed process or container with strict resource and environment controls. Stepflow's design prioritizes security, reproducibility, and platform independence.
• 🌐 Open, Portable Workflow Standard Build once, run anywhere. The Stepflow protocol is open and extensible, enabling workflow portability across different environments and platforms.

What Stepflow Enables

• Define AI workflows using YAML or JSON
• Execute workflows with built-in support for parallel execution
• Extend functionality through step services
• Handle errors at both flow and step levels
• Use as both a library and a service

Repository Structure

This repository contains multiple components:

• stepflow-rs/ - Main Rust-based execution engine and runtime
• stepflow-ui/ - Web-based frontend for workflow management
• sdks/python/ (stepflow-py) - Python SDK for building components
• sdks/typescript/ (stepflow-ts) - TypeScript SDK for building components

(back to top)

Getting Started

To get a local copy up and running quickly follow these simple steps.

Prerequisites

• Rust 1.70+ (for building from source)
• Python 3.8+ (for Python SDK examples)

Installation

1. Clone the repository

   git clone https://github.com/stepflow-ai/stepflow.git
   cd stepflow

2. Build the stepflow-rs project

   cd stepflow-rs
   cargo build --release

3. Run a sample workflow

   cargo run -- run --flow=examples/basic/workflow.yaml --input=examples/basic/input1.json --config=examples/basic/stepflow-config.yml

Contributing to StepFlow

Before contributing to StepFlow, you must sign our Individual Contributor License Agreement (ICLA):

python scripts/sign_icla.py

This is a one-time legal requirement that ensures clear licensing and intellectual property rights. The ICLA signature will be automatically verified when you submit pull requests.

For detailed development instructions, see CONTRIBUTING.md.

(back to top)

Usage

Quick Start Example

Here's a simple workflow that demonstrates basic Stepflow usage:

workflow.yaml:

input_schema:
  type: object
  properties:
    m:
      type: integer
    n:
      type: integer

steps:
  - id: add_numbers
    component: /python/add
    args:
      a: { $from: $input, path: m }
      b: { $from: $input, path: n }

outputs:
  result: { $from: add_numbers, path: result }

input.json:

{
  "m": 8,
  "n": 5
}

Run the workflow: This assumes workflow.yaml and input.json are in the current directory.

cd stepflow-rs
cargo run -- run --flow=workflow.yaml --input=input.json

Configuration

Create a stepflow-config.yaml file to define available plugins and routes:

plugins:
  builtin:
    type: builtin
  python:
    type: stepflow
    transport: stdio
    command: uv
    args: ["--project", "../sdks/python", "run", "stepflow_py"]

routes:
  "/python/{*component}":
    - plugin: python
  "/{*component}":
    - plugin: builtin

For more examples, please refer to the Documentation

(back to top)

Roadmap

• Serverless component execution: Run components in environments like AWS Fargate and Beam.cloud
• Flow-generating agent loops: Agent generates flows rather than single tool invocations.
• Integration with AI tools: Provide a runtime for popular frameworks like Langflow, Flowise, and n8n
• Richer component library: Pre-built components for common tasks
• Tracing and debugging: Enhanced tools for tracing workflow execution and debugging issues

See the open issues for a full list of proposed features (and known issues).

(back to top)

Contributing

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again!

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'feat: Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

For detailed development instructions, see CONTRIBUTING.md.

Top contributors:

(back to top)

License

Distributed under the Apache License. See LICENSE.txt for more information.

(back to top)

Contact

Project Link: https://github.com/stepflow-ai/stepflow

(back to top)