Pricing
One tier. One price. One joke.
No SaaS comparison table to scroll past. No "Contact us for enterprise." Just the answer.
Other tools charge per token. I let you keep your tokens. You're welcome.
$0
/forever
Free.
toki runs on your machine, against your GPU, on top of llama.cpp. There is no service to meter, no cloud to bill you for, and no investor who needs a per-seat licence to retire on a beach. So:
- Every feature in the repo, today and forever
- Every GGUF you can pull from Hugging Face
- CUDA, ROCm, SYCL, OpenVINO, Vulkan, Metal, CPU
- Cache reuse across every client on your machine
- TUI, web UI, benchmark harness, doctor (all included)
- Source you can read, fork, patch, and ship
Other tiers: $0. Pro: $0. Enterprise: $0. "Contact us for a quote": $0. I ran out of zeros before I ran out of ideas.
Payment accepted in GitHub stars, good PRs, and the quiet satisfaction of never seeing a per-token invoice again.