cookbooks

Cookbooks for Step 3.5 Flash

Welcome to the Step 3.5 Flash Cookbooks! This directory contains a collection of practical examples, recipes, and guides to help you get the most out of the Step 3.5 Flash reasoning model.

What is a Cookbook?

A cookbook is a self-contained project, integration guide, or example that demonstrates how to solve a specific problem or implement a feature using Step 3.5 Flash.

Whether it's a simple API integration, a guide on connecting with frameworks like LangChain/LlamaIndex, a complex reasoning agent, or a data processing pipeline, cookbooks are designed to be hands-on and easy to follow. We encourage you to add Integration Guides here as well, as they serve as excellent "recipes" for developers connecting Step 3.5 Flash with their existing tools.

Available Cookbooks

Here are the currently available cookbooks and integration guides:

• Hybrid Local Agent on MacOS
  Build a privacy-first, local agentic sandbox on MacOS. This guide demonstrates a hybrid architecture using Step 3.5 Flash for high-level reasoning and Qwen 2.5 Coder for high-volume tasks.

• OpenClaw Integration (Recommended) The recommended agent platform for Step 3.5 Flash. Learn to install, configure, and deploy OpenClaw for a seamless, powerful agentic experience.

• Roo Code Integration
  Configure Step 3.5 Flash as the backend for Roo Code (VS Code extension). Combines Flash's reasoning capabilities with Roo Code's autonomous coding features.

• Claude Code Best Practices Master Claude Code with Step 3.5 Flash. Covers environment setup, CLAUDE.md configuration, MCP integration, and Sub-agents for an optimized workflow.

How to Add a Cookbook

We welcome contributions! If you've built something cool with Step 3.5 Flash, please share it with the community.

1. Create a New Directory: Inside the cookbooks folder, create a new directory named after your topic. Use kebab-case (e.g., advanced-reasoning-agent, rag-implementation).
2. Add Your Code: Include all necessary source code. Python scripts (.py) or Jupyter Notebooks (.ipynb) are preferred.
3. Add Dependencies: Create a requirements.txt file listing all the Python libraries required to run your code.
4. Write Documentation: Create a README.md inside your directory explaining what the cookbook does and how to run it.
5. Submit a Pull Request: Push your changes to a branch and open a PR against the main repository.

Call for Contributions

We deeply value our community's creativity! If your cookbook is accepted and merged, we'd love to send you a small gift as a token of our appreciation.

• Contact Us: Please email us at developer@stepfun.com.
• Include: A link to your merged PR and your mailing address so we can ship your gift!

Basic Structure for a Cookbook

To ensure consistency, please follow this standard directory structure:

cookbooks/
└── your-cookbook-name/
    ├── README.md           # Documentation for your specific cookbook
    ├── requirements.txt    # Python dependencies
    ├── main.py             # Main entry point (or .ipynb)
    └── data/               # (Optional) Sample data needed for the example

Content of README.md (for your cookbook)

Your cookbook's README should ideally include:

• Title: Clear and descriptive.
• Description: What problem does this solve?
• Prerequisites: API keys, specific OS, etc.
• Installation: pip install -r requirements.txt
• Usage: Command to run the example.
• Example Output: What users should expect to see.

AI Tools for Documentation Refinement

To ensure high-quality documentation, we recommend using AI tools to refine your writing before submitting. Here are some prompts you can use to polish your README.md and comments.

Prompt 1: Structure and Clarity

"I have written a draft README for a code example. Please review it for structure and clarity. Ensure it has a clear Introduction, Prerequisites, Installation steps, and Usage guide. Suggest improvements to make it more professional and easy to follow for developers."

Prompt 2: Tone and Word Usage

"Reword the following technical documentation to be more concise and professional. Use active voice where possible. Ensure the tone is helpful and encouraging but technically precise. [Insert your text here]"

Prompt 3: Code Explanation

"Here is a Python function from my project. Please generate a clear, step-by-step explanation of how it works, suitable for inclusion in a tutorial or specific documentation. [Insert code here]"