YutoCode Docs

YutoCode is a next-generation AI coding agent built to break platform lock-in. It gives you one terminal-native workflow for multi-model, multi-account, and multi-platform engineering collaboration, with native compatibility for the OpenAI, Claude, Gemini, and OpenClaw ecosystems.

It is well-suited for reading repositories, changing code, running engineering commands, explaining project structure, and pushing tasks forward. It is also useful when you want to collapse previously scattered development actions into one continuous conversation.

If you have never used the terminal as a long-running working environment before, this documentation is still a good place to start.

Recommended Reading Order

1. Start with Installation to install the tool and verify the version.
2. Then read Quick Start to complete provider login and finish the first session.
3. Once the basics are familiar, read Common Commands and move the high-frequency commands into your daily workflow.
4. If login, permissions, or provider errors appear, go straight to Troubleshooting.

What YutoCode Is Good At

• explaining code structure, changing files, and running commands inside the current repository
• taking a natural-language request and drafting, refactoring, or adding tests around it
• handling quick analysis, search, debugging, and task framing before you even open the editor
• keeping “read code, change code, run commands, summarize results” inside one conversation
• preserving one usage style across different models and platforms instead of relearning every new product

What to Know Before You Start

• the public release currently targets macOS and Linux
• the primary command entry is yuto
• the product is positioned as natively compatible with OpenAI, Claude, Gemini, and OpenClaw
• the quick-start path focuses on the supported login flows that are already organized for first-time use

Quick Entry Points

• Installation
• Quick Start
• Common Commands
• Troubleshooting

Typical Working Rhythm

1. Ask Directly

Inside a project directory, run:

yuto

Once you enter the interactive session, describe the task the same way you would talk to a code-aware partner.

2. Run Once and Exit

If you only want one result and do not need a persistent conversation:

yuto --print "Explain this repository."

This is useful for quick summaries, one-off explanations, or temporary analysis.

3. Work Continuously Around a Project

The value of YutoCode becomes much more obvious once you let it read code, change files, and run commands. Its advantage is not merely “answering questions.” It is about keeping cross-platform model access, context understanding, and engineering execution in one continuous workflow.

A stronger prompt usually looks like:

Please inspect the entry structure of this repository first, then tell me where the login flow should be changed.

Instead of:

Help me change this.

The clearer the task, the more stable the output.