Naggy's brain
You have three options
to power Naggy.
Every option runs the same Naggy. They differ in where the thinking happens, who pays for it, and how private the conversation stays. Most people pick Cloud and never look back. Pick whichever fits.
Default for new users
Zero setup
Naggler Cloud
Ready the moment you install. Works out of the box. We cover the compute, served through TokenRouter — a third-party AI gateway running Claude-family models.
- Setup: None. It just works.
- Cost to you: Free during beta.
- Where thinking happens: Our relay, routed to TokenRouter — a third-party AI gateway serving Claude models.
- Powered by: Claude-family models via TokenRouter, a third-party AI gateway.
Your key, your billing
Bring Your Own Key
Drop in your own API key and Naggy uses that instead. Unlimited use, billed directly to you. Handy if you already have an Anthropic, OpenAI, or OpenRouter account, or want full control over the spend.
- Setup: Paste a key in Settings → Brain.
- Cost to you: Whatever the provider charges — typically pennies per day of active use.
- Where thinking happens: The provider's servers. We never see it.
- Compatible with: Anthropic / OpenAI / OpenRouter keys today.
Private by default
Local and Private with Ollama
One-click install flow inside the app. Your conversations never leave your machine — not to us, not to anyone. Perfect if you work on sensitive material or just prefer the old-school magic of running your own model.
- Setup: Click "Install Ollama" in Settings. The app handles the rest.
- Cost to you: Zero. You're using your own CPU/GPU.
- Where thinking happens: On your machine. Air-gapped capable.
- Trade-off: Slightly slower and a bit less sharp than Cloud on older laptops.