Speech, image & video
Transcription, text-to-speech and diffusion pipelines.
Speech and image tools split cleanly by hardware. Transcription and text-to-speech — Whisper and Piper — run well on CPU and are available now. Diffusion tools such as ComfyUI need a GPU to be worth using, so they are listed on request rather than sold as CPU plans.
Transcription is the quiet success story of self-hosted AI. Optimised Whisper builds turn a CPU instance into a transcription API that costs a flat monthly fee instead of per minute, and for meeting recordings and call archives — material most companies would rather not send to a vendor at all — that combination is hard to argue with.
Image and video generation are a different matter. Diffusion models are GPU-bound, and a CPU plan running ComfyUI would be a bad purchase dressed up as a cheap one. We list those tools so you know where we stand, and we will open them when GPU capacity can be priced honestly.
Speech, image & video we host
Whisper
Private speech-to-text behind your own transcription API.
from $12.99/moDetails →
Piper TTS
Fast neural text-to-speech that runs comfortably on CPU.
from $5.99/moDetails →
ComfyUI
Node-based image and video generation — waiting on GPU capacity.
on requestDetails →
Stable Diffusion WebUI
The AUTOMATIC1111 interface — GPU-bound, listed for completeness.
on requestDetails →
InvokeAI
A studio-style diffusion workspace — GPU-bound.
on requestDetails →Choosing between these
Straight answers
Is CPU transcription fast enough?
When will image generation be available?
Put your AI stack on your own box
Pick an app, pick a size, and have it running today. Month to month, cancel whenever.