AI models
Local (Ollama)
Open models such as Llama, running on your own server. No per-use cost; slow without a GPU.
In testingBuilt and tested against recorded API behaviour; not yet switched on for real accounts.
What airepeat uses it for
- Plans and writes your sites' code in AI tasks
- Answers in Create with AI and the planner
- Counts towards your workspace's AI budget
In testing: built and tested, but not yet run with real accounts.
Where it stands
Built in and tested; not yet run against a real account.
The key to create
No key: point airepeat at the address of an Ollama server you run. Models run on your own hardware, with no per-use cost; they are slow without a GPU.
Give each key the smallest access that works. Keys are encrypted at rest, only decrypted for the job that needs them, and never shown again after you save them.
Run your next site on your own servers
Free while in beta. Bring your own keys; you pay your providers directly.