stepfun-ai/Step-Code

Step Code

English · 简体中文

Swift execution, long-horizon reliability, and high token efficiency.

---

Step Code runs in your terminal and handles the full task loop—reading code, making changes, and running tests\. It works with the Step provider and discovers the available Step models after sign\-in\. MCP servers, Agent Skills, plugins, and multi\-agent orchestration work out of the box; you can delegate long\-running tasks to /goal and let Step Code drive them forward autonomously\.

Step Code comes with built-in StepPage publishing. Once your local page is ready, you can publish it as an accessible static website with a single command—bringing development, debugging, and delivery all within the same terminal\.

Why Step Code

• Token efficiency — tuned alongside Step models to consume fewer tokens for the same task; Long\-horizon tasks are split across parallel subagents, each with its own isolated context, so redundant content never enters the main conversation\.
• Static site publishing — the built\-in steppage plugin publishes a local directory to a shareable static URL in one command, with version management and rollback\.
• Long\-running task delegation — /goal hands an objective to Step Code, which works toward it autonomously, and /cron runs on a schedule; the status line shows a live timer for the active task\.
• Step provider access — sign in with a Step Plan account or use a Step Platform API key; switch among the models returned by Step in /model\.
• Ecosystem compatibility — MCP servers and Agent Skills work out of the box, and most Claude Code plugins are directly compatible\.

Installation

Pick the official installer for your platform\. It downloads the latest release, verifies the checksum, installs step to ~/.stepcode/bin, and updates your PATH; reopen your terminal and you are done\.

macOS / Linux / WSL

curl -fsSL https://static-openapi.stepfun.com/stepcode/install.sh | bash

Windows \(PowerShell\)

irm https://static-openapi.stepfun.com/stepcode/install.ps1 | iex

• Windows PowerShell support is currently in beta; on Windows, we recommend installing in WSL
• Installer flags: --version installs a specific version; --install-dir overrides the install directory
• Verify the installation in a fresh terminal:

step --version
step --help

Upgrade later with step update\.

Sign in to the Step provider

The default Step Code entrypoint exposes one built-in model provider: Step (StepFun). The profiles below select the Step service region and billing method; they are not separate providers.

| Profile | Credential | Billing | | -------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- | ------------------------------------------------- | | Step Plan (CN, platform.stepfun.com) | Browser OAuth; credentials refresh automatically | Usage included with Mini / Plus / Pro / Max plans | | Step Plan Oversea (platform.stepfun.ai) | Browser OAuth; credentials refresh automatically | Usage included with Mini / Plus / Pro / Max plans | | Step Platform (CN) | API key from platform.stepfun.com/interface-key | Pay per request | | Step Platform Oversea | API key from platform.stepfun.ai/interface-key | Pay per request |

Inside the TUI, /login opens the Step sign-in flow; from a shell, run step login. For API-key access, set STEP_API_KEY=<your_step_api_key> or enter the key in /login; headless runs can also pass --api-key <your_step_api_key>.

/status shows the current session and model status. Use step login status (or step login status --json) to check the Step credential and its validity. Stored credentials live in ~/.stepcode/auth.json with mode 0600; /logout clears the Step credential. After sign-in, /model lists the models discovered from the active Step endpoint.

Getting started

Change into your project directory:

cd /path/to/your/project
step

Describe the task in the TUI, or submit one at launch:

step -p "What is the tech stack of this project? What is each directory responsible for? How do I run it locally?"

Run /init to generate an AGENTS.md project guide; an existing CLAUDE.md is used as\-is\.

| Entry | Command | Use case | | --------------- | --------------- | -------------------------------------------------------- | | Interactive TUI | step | Exploring code, ongoing conversations, reviewing changes | | Headless | step -p "..." | Scripts, CI, batch jobs |

Examples

Start the TUI in any project and enter:

What is the tech stack of this project? What is each directory responsible for? How do I run it locally? Don't modify any files yet.

Then one that makes changes:

Create a reference table for the error codes in src/api/errors.ts (code, meaning, trigger scenario) and save it to docs/errors.md

Resume a session

step -c   # Continue the most recent session
step -r   # Browse session history and pick one to resume

Inside the TUI, /resume finds past sessions and /hotkeys lists all shortcuts\.

Common shortcuts

| Key | Action | | -------------------------------------- | ---------------------------------------------------------------------------------- | | Enter | Send message | | Shift+Enter / Ctrl+J / Alt+Enter | Newline | | Enter | Queue a message while streaming \(delivered after the current tool call finishes\) | | @ | Reference a project file | | !command | Run a shell command and hand the output to the model | | Shift+Tab | Cycle permission mode | | Ctrl+O | Toggle tool output | | Esc | Interrupt the running task |

Migrate from other agents

• MCP config: Claude Code / Codex MCP configuration is imported automatically on first launch; source files are never modified, the import is idempotent, and secrets are never inlined\.
• Plugins: most Claude Code plugins work as\-is, managed via /plugin\.
• Project guides: an existing CLAUDE.md is used as\-is; /init generates the equivalent AGENTS.md\.

Build from source

Developing Step Code or running from this repository requires Git, Node\.js, and pnpm:

git clone https://github.com/stepfun-ai/Step-Code.git
cd Step-Code
pnpm install --ignore-scripts
pnpm run build
pnpm step

When running from the source tree, replace step with pnpm step\.

Security

• Four permission modes \(Ask / Read Only / Bypass / Autopilot\), cycled with Shift+Tab; dangerous commands require a separate confirmation dialog in every mode\.

Uninstall

Make sure no step sessions are running before you uninstall\.

Remove the binary: delete the step executable from the install directory \(default ~/.stepcode/bin\)\. The installer appends a PATH block marked # stepcode to your shell config; removing that block completes the uninstall \(zsh: ~/.zshrc, bash: ~/.bashrc, fish: the fish_add_path config\)\.

Remove user data \(optional\): everything lives under ~/.stepcode \(config, credentials, sessions, logs\); delete the directory to remove all Step Code data:

rm -rf ~/.stepcode

Docs \& community

• Quick start · Interaction \& input · Platforms \& models · All docs \(links finalized at release\)
• Contributing · \[Report a bug / request a feature\]\(issues link added at release\) · Report a security issue
• When filing an issue, include step --version, reproduction steps, and how you run Step Code; strip credentials and private project content first\.

License

Released under the MIT license; see LICENSE\.