Rayucode brings the full power of the Rayu AI coding agent into Visual Studio Code.
Powered by the same engine, tool platform, and slash commands as the Rayu CLI, Rayucode runs as a dedicated assistant in your sidebar, allowing you to build, debug, and review code changes seamlessly inside your editor.
Features
Multi-Provider AI: Use Claude (Anthropic), OpenAI, DeepSeek, Google Gemini, AWS Bedrock, or custom OpenAI-compatible models.
Interactive Review Card: Review files changed by Rayu turn-by-turn with side-by-side git diffs, and keep or undo changes individually or in bulk.
Slash Commands & Palettes: Type / to trigger the slash command palette (/help, /compact, /cost, /model, /clear, /review), and @ to mention workspace files.
Permission Modes: Control execution safety with one-click permission cycling (Shift+Tab) across Plan, Default, Accept Edits, and Full Access.
Context Usage: Live context window tracking in the panel header.
MCP Server Integration: Manage and reconnect Model Context Protocol servers directly from the extension.
Single Credential: Shares configuration and credentials with the Rayu CLI (~/.rayu/).
Getting Started
Install the Rayucode extension in VS Code.
Open the Rayucode icon in the Activity Bar.
Sign in to Rayu or connect with an API key (/connect in the terminal).
Start chatting and coding!
Remote-SSH Limitation
Account sign-in via OAuth loopback requires an interactive local browser and is not currently supported in Remote-SSH sessions. For Remote-SSH workspaces, configure an API key via /connect in your remote terminal: Rayucode detects and loads shared credentials automatically.
