Quickstart

Tokens.Pay gives you one prepaid key for 16 models. The API is OpenAI-compatible: keep your SDK, change the base URL and the key.

1. Get a key

  1. Sign in with a Gridheap Account and add credit: any amount from $5 to $500, plus a 5% fee (minimum $0.79).
  2. Copy your key when it is shown, right after payment. It starts with rk_ and is shown once; store it in your password manager or secret store.
  3. Export it for the samples on these pages: export TOKENS_PAY_KEY="rk_…"

2. Point your SDK at the base URL

Base URL
https://api.airuntime.dev/v1
Header
Authorization: Bearer rk_…
Model
A model id from the model list, such as deepseek-flash

Your code calls AI Runtime at api.airuntime.dev directly; requests never pass through Tokens.Pay.

3. Make your first call

curl

curl https://api.airuntime.dev/v1/chat/completions \
  -H "Authorization: Bearer $TOKENS_PAY_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-flash",
    "messages": [{"role": "user", "content": "Say hello in five words."}]
  }'

The answer is a standard chat completion. It also carries airuntime.cost_usd_micros: what this request cost, in millionths of a dollar. At DeepSeek V4.1 Flash’s list price that is $0.15 per 1M input tokens and $0.60 per 1M output tokens.

Next

Get a key, then make your first call.

From $5, plus a 5% top-up fee (minimum $0.79). No subscription; paid credit never expires.

Add credit