Yishan

English · 简体中文

Yishan helps agents know your projects better.

Spend less time explaining. More time deciding.

Why Yishan?

Every time you start a new agent session, you repeat the same context — project structure, conventions, decisions already made. The burden of explanation falls on you, every time.

Yishan gives agents persistent project memory, shared context, and repeatable workflows. They find what they need on their own. You stop repeating yourself.

Core Features

• Isolated workspaces — each task gets its own branch, terminal, and file state. Switch between tasks without losing anything.
• Memory that grows with your project — .my-context and built-in memory accumulate decisions, notes, and session summaries automatically. Agents find context on their own.
• Skills that remove the ceremony — define how you work once — review patterns, test rules, handoff steps. Every agent follows the same playbook, every time.

More

• Autopilot — schedule recurring agent jobs on a cadence — code reviews, health checks, weekly summaries — running automatically while you focus elsewhere.
• PR visibility — see pull request state, checks, and review status without leaving Yishan.
• Works with your tools — open files in VS Code, Cursor, JetBrains, Sublime, Xcode, and more. Yishan bridges to your tools instead of replacing them.
• Bring your own agent — works with OpenCode, Claude, Codex, Gemini, Cursor, and Pi. Use the agents you already know.

Roadmap

• Remote host workspaces — run workstreams on a remote machine (in progress)
• CLI + agent workflow integration — spawn multiple workspaces from chat or CLI (in progress)
• Development lifecycle management — issue to workspace to PR in one flow (planned)
• Mobile remote control — monitor and steer workspaces from your phone (planned)

Getting Started

Download for macOS at yishan.io.

To build from source and contribute, see CONTRIBUTING.md.

Links

• Website
• Changelog
• Contributing
• License