Skip to main content
Roo Code is a Coding Agent plugin that runs in an IDE and can complete development tasks such as code generation, debugging, and file modification through natural language. By configuring the OpenAI Compatible API, Roo Code can connect to Step models and directly call inference capabilities inside the IDE. This document explains the prerequisites, configuration method, parameter recommendations, and common issues for integrating the Step API with Roo Code.

​
Overview

Roo Code is suitable for multi-turn code collaboration in the editor. After configuration is complete, you can directly call Step models in the plugin panel to complete coding tasks.

​
Prerequisites

​
Development Environment

The following IDEs are supported:
  • Cursor
  • VS Code

​
Install the Roo Code Plugin

Search for and install the following in the IDE Marketplace: 
Roo Code
After installation, the following entries usually appear in the IDE: 
Roo: Open Chat
Roo: New Task
Roo: Settings

​
Subscribe to Step Plan

Before you start configuring, make sure your current account has an active Step Plan subscription. Only accounts with the required plan or calling permissions can use model calls and quotas normally. To subscribe or purchase, visit Step Plan subscription

​
Obtain a Step API Key

Visit the Step Open Platform Console to create an API Key and keep it safe:

​
Configuration Steps

Open the Roo Code settings page and configure the API Provider.

​
Connection Configuration

Set the following parameters:
  • API Provider: OpenAI Compatible
  • Base URL: https://api.stepfun.ai/step_plan/v1
  • API Key: enter the Step API Key
  • Model ID: enter <model_id>
Note: In the examples below, replace <model_id> with either step-3.5-flash-2603 or step-3.5-flash.

​
Parameter Recommendations

The recommended configuration is as follows: 
Stream: Enabled
Context Window: 256000
Include max output tokens: Off
Max Output Tokens: -1

​
Compatibility Notes

In some plugin versions, if Include max output tokens is enabled, the request may automatically include the following fields: 
max_tokens
max_output_tokens
Different OpenAI Compatible APIs may differ in compatibility with these fields, which may cause the request to fail and display: 
OpenAI completion error: Connection error
If you encounter this issue, you can try:
  • Turn off Include max output tokens
  • Or keep Max Output Tokens = -1

​
Test the Integration

The original material does not list test steps separately, but after configuration is complete, you can launch a minimal task to verify whether the model invocation is successful, such as creating a new conversation and trying to generate simple code.

​
Common Issues

​
Connection error

If the following error occurs: 
OpenAI completion error: Connection error
Please check:
  1. Whether the Base URL is https://api.stepfun.ai/step_plan/v1.
  2. Whether the API Key is valid.
  3. Whether the Model ID is <model_id>.
  4. Whether the Context Window is set to a sufficiently large value (recommended 256000).

​
API Key Error

If the following error occurs: 
401 Incorrect API key
Please check:
  • Whether the key was copied completely
  • Whether it belongs to the correct Step environment

​
Summary

After completing the OpenAI Compatible configuration, Roo Code can call Step models in the IDE for code generation, file editing, and task collaboration. It is recommended to prioritize the recommended parameters and first complete a minimal task verification.