Contact Us
Questions about Vidman AI? Comparing models, sizing a workload, or puzzled by a bill? Write to us and we will sort it out.
Send Us a Message
Fill out the form below. We reply within 24 hours on business days.
Frequently Asked Questions
Quick answers to common questions.
How does pricing work?
Inference carries published per-million-token input and output rates — the full rate card sits in the model catalog. Fine-tuning runs across dedicated GPUs metered by the second. The public calculator prices any workload before a rupee is spent.
Where does my data run?
Inside India, on an Indian-operated stack — every inference call, every fine-tuning and training run, and every dedicated endpoint. Nothing is retained anywhere: prompts and responses are processed in memory and discarded when the request completes.
What models does Vidman AI serve?
Frontier open and closed models behind one OpenAI-compatible endpoint — Claude, DeepSeek, Gemma, GLM, Kimi, MiMo, MiniMax, Nemotron and Qwen. The model catalog lists each with its context window and rates.
What is Vidman AI Adaptive?
A single endpoint that hands each request to the best model for that task at the lowest cost, so your code quits hard-coding model ids. Same request shape, same published rates — routing never changes what a token costs.
Do I own my fine-tuned model?
Yes. Fine-tuning runs across dedicated GPUs metered by the second, and the trained weights are yours outright — exported, portable, runnable anywhere.
Is the API OpenAI-compatible?
Yes. Aim your existing OpenAI SDK at https://api.vidman.ai/v1 with your Vidman AI key and the code runs as-is. The docs’ API overview covers authentication and examples.
When will I hear back?
Within 24 hours on business days. The team sits in Bangalore, India and works Indian Standard Time, Monday to Friday, 9am to 6pm IST.