Services
Small calculators for the GPU bill. Each one uses a published hourly price and a published throughput, and it shows the source. Four are live: OCR, embeddings, speech-to-text, and open LLM hosting.
Pages per month in, a GPU and an open OCR model out, with the monthly bill and the break-even against Textract, Document AI, Azure, and Mistral.
Open the calculatorCorpus size and re-embed frequency in, an open embedding model and a GPU out, with MTEB scores next to OpenAI, Voyage, Cohere, and Google prices.
Open the calculatorAudio hours in, Whisper, Parakeet, or Canary on a GPU out, with leaderboard WER next to OpenAI, Deepgram, AssemblyAI, Google, and AWS prices.
Open the calculatorTokens and concurrent users in, VRAM, GPU count, and a monthly bill out for Llama, Qwen, gpt-oss, and more, next to API token prices.
Open the calculatorFree, from this page
One email with the deployment checklist and the link to price your own page count. The calculator is where the model, GPU, and monthly bill get filled in.
OCR deployment plan
No monthly figure is shown here, because this page has no page count yet.