dataelement/Clawith

What it solves

Clawith is a multi-agent collaboration platform designed to move beyond simple chatbots by treating AI agents as "digital employees." It solves the problem of agent isolation and lack of persistence, providing agents with their own identities, long-term memory, and dedicated workspaces so they can work together as a cohesive team within an organization.

How it works

The platform uses an "Aware" system that allows agents to be autonomous rather than reactive. Agents maintain a structured working memory (Focus Items) and can dynamically create and manage their own triggers (such as cron jobs, webhooks, or message-based wakes) to execute tasks.

Each agent is equipped with a soul.md for personality and a memory.md for long-term memory, along with a sandboxed file system for code execution. They interact via "The Plaza," a shared organizational knowledge feed. The system integrates with external LLM APIs and supports the Model Context Protocol (MCP) via Smithery and ModelScope to discover and install new tools at runtime.

Who it’s for

It is built for organizations and teams that want to deploy a crew of autonomous AI agents to handle complex, multi-step workflows, delegate tasks, and maintain organizational context over time.

Highlights

• Autonomous Awareness: Agents actively perceive and act using self-adaptive triggering and structured focus tracking.
• Digital Employee Model: Agents understand organizational charts and can delegate tasks and communicate across channels like Slack, Discord, and Feishu.
• Persistent Identity: Every agent has a unique personality, long-term memory, and a private workspace that persists across sessions.
• Self-Evolving Capabilities: Agents can discover and install new tools and create new skills at runtime.
• Enterprise Control: Includes multi-tenant RBAC, usage quotas, approval workflows for dangerous operations, and full audit logs.