OpenAI-compatible
Keep the OpenAI SDK and the code you already have. Change the base URL, the key and the model name.
Four models behind one OpenAI-compatible API, running on Torki's own machines. Send text, images and video. Pay only for the tokens you use.
Pick a language. The request runs, the answer streams back, and it shows up in your console.
The same four modes as the Torki app. All four read up to 128K tokens of context. Point at a card to see what it costs and what it is for.
The pieces you need to build on Torki, in the console you sign in to with your Torki account.
Keep the OpenAI SDK and the code you already have. Change the base URL, the key and the model name.
Send an image by link or inline, and video up to 10 minutes long, to any of the four models. Every image and frame is screened.
Try a prompt on any model in the browser, watch it stream, then switch to Code to copy the request.
A dashboard, request logs, usage you can export as CSV, invoices, and spend limits for your team.
Your tier rises with the credit you buy, and so do your spend limit and requests per minute.
Torki's own model, on hardware Torki owns and operates. The text and image safety checks run there too.
No subscription. Sign in with your Torki account, add a $10 Starter pack, create a key and make your first call. Every call draws from your credit.
Also available in rupees at checkout