llms.txt git:20260418.07240ecA

Immutable. This exact content is served forever at /api/v1/blob/1fd521b4bf5ced16.

# Apra Fleet

> AI-managed fleet orchestration for Claude Code — run, update, and coordinate multiple Claude Code agents from a single hub.

Apra Fleet is a multi-agent orchestration layer that lets a PM agent delegate work to a fleet of Claude Code instances via MCP tools, Git, and SSH. Each fleet member runs its own Claude Code session; the PM agent controls lifecycle, skills, and task assignment. Members can run different LLM backends (Claude, Gemini, Codex, Copilot) and be mixed freely within a single fleet.

## Docs

- [Readme](README.md): Installation, configuration, member registration, and day-to-day usage for operators.
- [Vocabulary](docs/vocabulary.md): Shared terminology — member, task, skill, PM, fleet, doer/reviewer pattern.
- [Provider Matrix](docs/provider-matrix.md): Which LLM providers are supported, their capabilities, and how to configure them.
- [FAQ](docs/FAQ.md): Common questions about setup, troubleshooting, and the doer-reviewer loop.
- [Architecture](docs/architecture.md): How the fleet hub, MCP server, and members interact at a system level.