This project provides a bridge between Feishu/Lark messenger and a local Claude Code CLI, enabling users to interact with Claude directly from chat interfaces. It's designed for developers and power users seeking to integrate advanced AI code assistance into their existing communication workflows, offering real-time updates and persistent chat sessions within a controlled local environment.

How It Works

The bot forwards Feishu/Lark messages (DMs or group mentions) to a locally running claude CLI instance, operating within a user-specified working directory. It utilizes a real-time streaming card interface within Lark/Feishu for Claude's text and tool call updates, ensuring immediate feedback. Key features include per-chat session management for conversational continuity, preemptive handling of rapid messages, and support for multiple named workspaces, allowing users to switch between project contexts seamlessly. Images and files sent to the bot are downloaded and made accessible to the local Claude CLI.

Quick Start & Requirements

• Install: npm i -g lark-channel-bridge or pnpm add -g lark-channel-bridge.
• Prerequisites: Node.js >= 20, claude CLI installed and logged in (see Claude Code Quickstart), and a Lark/Feishu PersonalAgent app (can be created via an integrated QR-code wizard).
• First Run: Execute lark-channel-bridge run. This command initiates a QR code wizard in the terminal for binding the Lark app.

Highlighted Details

• Real-time streaming cards update Claude's responses and tool calls instantly within Lark/Feishu.
• Maintains distinct, persistent sessions for each chat, allowing conversations to resume where left off.
• Supports multiple named workspaces (/ws commands) with context-specific session tracking.
• Interactive cards with buttons enable actions like /help, /ws list, and /status.
• Handles direct uploads of images and files, making them available to the local Claude CLI.

Maintenance & Community

No specific details regarding maintainers, community channels (e.g., Discord, Slack), or roadmap are provided in the README.

Licensing & Compatibility

The project is released under the MIT License, which generally permits commercial use and integration into closed-source projects.

Limitations & Caveats

Running the bridge as a background daemon requires a global npm installation to prevent issues with cache garbage collection affecting the daemon's executable path. Access control configuration, while powerful, can be complex to manage, requiring careful handling of user and chat IDs to avoid accidental lockouts. Daemon service definitions (launchd, systemd, Task Scheduler) hard-code the bridge CLI path, necessitating manual updates or re-installation if the path changes. Lark's platform may route events randomly to one of multiple live WebSocket connections if the same app is started more than once.