Any NVIDIA alternatives?

[aguja262]

Hi, I'm new to this so I prefer to ask since I got stuck on a step in the process.
I created an NVIDIA account to get the API key, but they won't let me generate it because they are asking me to contact support to verify the account.
I did that, but I still haven't heard back. Maybe they just take a long time, but there are so many people in the forums asking for manual verifications even though the forums aren't meant for that, that I just don't know what to think.
I'd like to know if there is any alternative regarding that step.

[dgesteves]

You don't need NVIDIA. It's only the default in the Quick Start. FCC works with any provider in the README's provider catalog, and a few of them have a free tier you can sign up for with a normal account.

On the NVIDIA side, their account access post says to email help@build.nvidia.com with the email you registered, any phone numbers you tried, and a screenshot of the error. People in that thread report waiting weeks, so I wouldn't block on it.

Google AI Studio (Gemini)

1. Create a key at https://aistudio.google.com/apikey
2. Admin UI → Providers → Gemini → Add key, paste it, Save
3. Search the MODEL dropdown for gemini/models/gemini-3.1-flash-lite and click Apply

Google's pricing page lists that model as free of charge on the free tier, with rate limits. FCC's hint on that key also notes free-tier data may be used to improve Google's products outside the UK/CH/EEA/EU.

OpenRouter

1. Create a key at https://openrouter.ai/keys
2. Providers → OpenRouter → Add key, Save
3. Set MODEL to open_router/openrouter/free (or a tool-capable model whose ID ends in :free) and Apply

Free models are capped at 50 requests a day until you've bought $10 of credits in total, then 1000. Claude Code makes a request per tool step, so 50 goes quickly.

If your machine can run a model locally, Ollama or LM Studio need no account at all: start the server and pick an ollama/<tag> or lmstudio/<model-id> model.

Whichever you choose, make sure MODEL itself changes: it defaults to nvidia_nim/nvidia/nemotron-3-ultra-550b-a55b, so adding a key on its own isn't enough. The Admin UI saves these settings (GEMINI_API_KEY / OPENROUTER_API_KEY and MODEL) to ~/.fcc/.env. Start fcc-claude again afterwards.

I checked this against the README and source at v6.18.5. I haven't run it with real keys.

[aguja262]

thank you so much!!