nVidia API

· 1 min read

WELL SLAP MY ASS AND CALL ME GRANDMA

So nvidia awhile ago announced build.nvidia.com.

Wicked idea, FREE <-- yep FREE (for now) API Calls to advanced frontier models, llms, coding, image gen. Whole nine yards... well image gen is behind a purchase gate but the rest is in the free tier.

So.. of course , I wired this into Project HOME.

HOME now has multiple endpoint options to ensure that the chat never dies.
Ontop of that it runs in modes, depending on the mode depends on the model called.

If she's in her standard mode, she's using Qwen3.6 if she's coding , she's using DeepSeekv4 Flash.

If she's using advanced tool calls or needs to REALLY know what she's doing.
Nemotronsuperultra-550b (I ... reading that back... I dont think that's the model name but it's close enough)

This has been a major MAJOR improvement to the project.
The project now has legs and arms.. it has eyes, a mind, a subconscious even.

It has a place to deliberate and can actually fucking use it now! It's awesome.
The tool calls keep failing, however that's only on occasion.

Anyway - that's all for today

All the best, Be Your Best
- h4 (b'.')b

More from the blog