llms.txt Content
# SwarmClaw
> SwarmClaw is a self-hosted AI agent runtime and control plane: run agents across 25+ model providers and CLI backends, delegate work between agents, automate recurring tasks, manage long-term memory, and bridge agents into chat platforms.
SwarmClaw installs from npm (`@swarmclawai/swarmclaw`), runs a local server with a web UI, and keeps state on your own machine or VPS.
## Start Here
- [Home](https://www.swarmclaw.ai): Product overview, features, and install options
- [Downloads](https://www.swarmclaw.ai/downloads): Desktop app and install downloads
- [Documentation](https://www.swarmclaw.ai/docs): Full product documentation index
- [Getting Started](https://www.swarmclaw.ai/docs/getting-started): Install and run SwarmClaw with npm, pnpm, yarn, bun, or the install script
- [Release Notes](https://www.swarmclaw.ai/docs/release-notes): What changed in recent versions
## Agents & Orchestration
- [Agents](https://www.swarmclaw.ai/docs/agents): Build agents with personalities, system prompts, tools, and MCP integrations
- [Providers](https://www.swarmclaw.ai/docs/providers): 25+ built-in providers — Anthropic, OpenAI, Gemini, OpenRouter, Ollama, CLI backends, and custom endpoints
- [Delegation](https://www.swarmclaw.ai/docs/delegation): Agent-to-agent handoffs and delegation to Claude Code, Codex, Gemini CLI, and other CLI backends
- [Chatrooms](https://www.swarmclaw.ai/docs/chatrooms): Persistent multi-agent collaboration in a shared room thread
- [Structured Sessions](https://www.swarmclaw.ai/docs/structured-sessions): Bounded multi-agent runs with templates, artifacts, and operator controls
- [Tasks](https://www.swarmclaw.ai/docs/tasks): Kanban-style task board with dependencies, retries, and run history
- [Projects](https://www.swarmclaw.ai/docs/projects): Durable scope grouping agents, tasks, schedules, and goals
- [Scheduling](https://www.swarmclaw.ai/docs/scheduling): Cron-based schedules for agents, chats, and sessions
- [Tools](https://www