Honest Pricing
Ten models. One price. No PhD required.
Built by one person with no investors to please. Every plan gets all 14 products, all 172 geos, every language. Same AI, different credit limits.
Your daily driver. Ask anything, code anything.
- $40 monthly credits
- All models under 30B
- Code generation + voice
- 8K context
- Smart routing (auto)
The whole brain. See, think, reason, build.
- $150 monthly credits
- Everything in Starter
- Vision + deep reasoning
- Camera input
- 32K context window
For the person who maxes out everything.
- $300 monthly credits
- Everything in Pro
- Priority queue
- API access
- 128K extended context
- Team workspace (soon)
A GPU with your name on it. Literally.
- $750 monthly credits
- Everything in Business
- Dedicated NVIDIA B200 GPU
- 750B+ parameter models
- Zero wait — no throttle
- Custom model routing
Your credits go further than you think
Every query costs tokens. The smart router (toolkit-auto) quietly picks the cheapest model that nails your task — so you stop burning money on overkill.
Per-model token pricing
Per 1M tokens. Cached input = repeated context (system prompts, multi-turn history) at 90% off. The smart router optimizes automatically.
No feature walls. Seriously.
Same AI. Same products. Same 172 geos. Higher plans just unlock bigger models and more credits.
No board meetings. No growth targets. Just GPUs.
One person built this. No investors. No ads. No one telling me to make the free tier worse so you upgrade faster.
A student in Mumbai pays less than a trader in London, and both get the same models. I charge what GPUs cost, plus enough to keep the servers warm.