Providers and models
FreeGent spreads work over many free-tier and local models instead of depending on one. These pages cover where the models come from and how FreeGent switches between them.
| Page | Covers |
|---|---|
| Default providers | The providers FreeGent ships with, where to get keys, and the variable names |
| Priority, fallbacks and retries | The model priority list, cooldowns, rotation, retries and thinking level |
| How requests reach providers | The hosted proxy, the Worker, and when you need your own |
| Local models | Ollama, vLLM, LM Studio and any OpenAI-compatible server |
On freegent.ai the only setup is pasting an API key. Besides the default providers, any OpenAI-compatible endpoint can be added as a custom model.