Clawdi

**名称:** Clawdi **描述:** Clawdi is a cloud-based platform designed to host, manage, and coordinate multiple AI agents, their environments, sessions, memory, skills, and background tasks within a single unified workspace. It addresses the fragmentation problem where AI agents from different frameworks, projects, or devices operate in isolation, each with their own keys, memory stores, and skill sets that cannot interact or share context. The core offering of Clawdi is a centralized deployment system for AI agents. Users can instantiate new agents—such as Claude Code, OpenClaw, Hermes, or custom models—each running inside its own isolated environment. Every agent has its own dedicated session history, persistent memory, and a set of skills (functions or tools) that can be attached or removed. The platform also manages cron jobs for recurring tasks, external application connections (e.g., APIs, databases), and API keys required by each agent. Agents are organized into projects or groups, with the ability to set access boundaries. For example, a production agent like "openclaw-prod-01" can be walled off from certain other agents (e.g., Claude Code or Codex) to enforce security and data isolation. Sessions are tracked individually: each agent maintains a chronological log of all interactions, allowing users to resume work from any point. The interface shows real-time resource usage (CPU, RAM) and uptime for each agent instance. A key feature is memory management. Clawdi provides persistent storage that agents can read from and write to, enabling long-term context retention across sessions. Skills are modular units of capability—such as web scraping, file processing, or API calls—that can be attached to any agent. Users can define custom skills or import existing ones. Cron job