Can add ollama 1000 times per month

https://build.nvidia.com/models，This can also be added

I'd like to to take this issue on

Thanks both. @zhangyu1324 NVIDIA Build (build.nvidia.com), already in the catalog, currently enabled = 0. It's the same backend as integrate.api.nvidia.com/v1, which we evaluated previously: ~$5 of signup credits that exhaust and then require a card to top up. That fails the "recurring free quota with no card" criterion (same reason DeepSeek, Together, Fireworks were rejected). The row is preserved so it can flip back on if NVIDIA changes pricing, see the disable note in server/src/db/index.ts.

Ollama Cloud is a real lead. Their pricing page shows a Free plan ($0/mo, no card) with cloud hosted models, OpenAI-compatible API, 1 concurrent model and 5-hour session limits. Quota shape is GPU time rather than monthly tokens but it's recurring and self serve, so we should add it. Please make a PR and I'll review and merge @LaloLalo1999 @zhangyu1324