zts212653/clowder-ai

What it solves

Clowder AI addresses the friction of managing multiple AI models (like Claude, GPT, and Gemini) manually. Instead of the user acting as a manual router—copying and pasting context between different chat interfaces—Clowder provides a platform layer that allows these isolated agents to work together as a cohesive team.

How it works

Clowder acts as an orchestration layer that sits above existing agent CLIs (such as Claude Code, Codex CLI, and Gemini CLI). It provides a unified infrastructure for:

• Agent-to-Agent (A2A) Communication: Using @mention routing and structured handoffs for asynchronous messaging.
• Persistent Identity: Maintaining roles, personalities, and memories for agents across different sessions.
• Shared Memory: A centralized store for evidence, decision logs, and lessons learned.
• Cross-Model Review: Enabling one model to produce work (e.g., code) and another to review it natively.
• Plugin Framework: Integrating MCP tools and IM adapters for platforms like Telegram and Feishu.

Who it’s for

It is designed for operators who use multiple LLM-based agents for complex tasks and want to automate the collaboration, handoffs, and knowledge sharing between them.

Highlights

• Multi-Agent Orchestration: Routes tasks to the specific model best suited for the job (e.g., Claude for architecture, GPT for review).
• A2A Messaging: Built-in asynchronous communication between agents.
• Shared Knowledge Base: Persistent memory that prevents loss of context across sessions.
• Broad Compatibility: Supports Claude Code, Codex CLI, Gemini CLI, and opencode.