> ## Documentation Index
> Fetch the complete documentation index at: https://docs.u-kiyo.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Quickstart

> Choose an offer, set a budget, and connect after provisioning.

1. Open [@UkiyoComputeBot](https://t.me/UkiyoComputeBot) in a private Telegram chat.
2. Send `/rent` or choose **Rent a GPU**.
3. Choose an available GPU option. Check quantity and VRAM per GPU.
4. Enter a budget in USD within the range shown. The minimum is \$5; the maximum depends on the offer and available duration.
5. Review the hourly price, compute credit, and estimated runtime.
6. Continue to the secure payment page. Supply your receipt/access email if it is not already associated with the checkout. Agent checkout requires the email saved during token onboarding.
7. Complete payment. Wait for provisioning to finish; do not pay twice while waiting.
8. When ready, use the access details in your email or **My GPUs**. See [SSH](/guides/ssh).

The GPU-ready email combines order details and access information, with a receipt link when available. Delivery and provisioning are not instantaneous.

## API tokens are optional for Telegram rentals

You only need a token for the API, CLI, MCP, or Skill. In a private chat, send `/tokens`, choose **Create token**, and enter an email if your account does not have one. Once saved, that email is not requested again for subsequent token creation. Copy the secret immediately; it is shown once.

Finish or cancel a pending rental budget entry before opening token management. Token onboarding must not consume your budget reply.

If payment succeeds but access is missing, check junk mail and deployment status, then [contact support](/essentials/support) with the order or deployment ID.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.