Locally hosted LLMs are free, with no per-token billing. Upgrade to Pro for cloud models.
All available non-cloud LLMs, including local Ollama models, are free. Cloud models require Pro.
Sub-agents, code verification, parallel panes, and design mode: the key features are built in, not bolted on.
Ships with expert playbooks for systematic debugging, safe refactoring, secure coding, test writing, and more. The agent pulls the right one into context exactly when a task needs it.
Delegate work to focused specialists: one scans the repo, another audits the diff, and the main thread stays fast and focused.
Before it calls a task done, Fantasy runs your test suite and has a strong reviewer audit every changed file against what you actually asked. No silent "it compiles, ship it."
Fantasy understands the formats your tools already use: MCP servers, skills, memory, and rules carry over. Bring your whole setup across with no lock-in and nothing to rewrite.
Run several agent conversations at once: tile them side by side, switch between projects, or pop panes into separate windows. Built-in i3-style panes and tabs make real multitasking feel natural.
Click any element in the live preview and tell the agent what to change. It edits your real source code: components, styles, and layout, not a disposable mockup.
Copy, paste, and you're ready to go.
$ powershell -ExecutionPolicy ByPass -c " irm https://fantasyai.cloud/install.ps1 | iex "Download Fantasy Desktop and give your projects an AI copilot that runs on your machine.
Download for Windows