>_ FAQ

Fermata is a native macOS application for orchestrating multiple Claude Code agents in parallel. It provides a visual canvas where each agent is a node — you spawn them, review their work, approve tool use, and merge their changes.

Yes. Fermata is free during the beta period. We plan to introduce a Pro tier in the future for features like mobile companion apps and cloud sync, but the core orchestration features will remain free.

Yes. Fermata is an orchestrator — it directs Claude Code agents, but the actual AI runs through Anthropic's Claude Code CLI. You need it installed and authenticated before using Fermata.

Fermata requires macOS 15 (Sequoia) or later, running on Apple Silicon (M1 or newer). For the best visual experience with Liquid Glass, macOS 26 (Tahoe) is recommended.

No. Fermata runs entirely on your machine. Your code, prompts, and agent conversations never leave your computer. The only network traffic is between the Claude CLI and Anthropic's API — which is handled by the CLI, not by Fermata.

SDD is Fermata's structured workflow for complex features. You write a spec describing what you want, AI generates a strategy, you annotate and refine it, then agent teams execute the tasks in parallel. It turns ideas into shipped code through a repeatable pipeline.

Claude Code runs one agent in one terminal. Fermata lets you run many agents simultaneously on a visual canvas, each in their own git worktree. You get parallel execution, visual status monitoring, structured approval flows, and SDD for complex features.

Not currently planned. Fermata is built with 100% native SwiftUI and is deeply integrated with macOS. We're focused on making the Mac experience excellent first.

Each agent session gets its own git branch and worktree automatically. This means agents work in isolation — they can't interfere with each other or your main branch. When an agent finishes, you review the diff and choose to merge, edit, or discard.

Not yet. Fermata currently works exclusively with Claude Code. Multi-provider support (Gemini, etc.) is on the roadmap.