OpenAI-compatible · your own GPUs

Sell inference from your hardware, metered by the token.

TokenKiln puts a prepaid, per-token gateway in front of your vLLM, Ollama, or TGI stack. Customers top up credits, mint API keys, and point any OpenAI SDK at your base URL.

Start selling tokens Read the docs

Prepaid credits

No invoices, no surprises

Exact metering

Usage from the engine, not estimates

SSE streaming

Token-by-token passthrough

Spend controls

Per-key monthly limits

Models & pricing

View all →

Model	ID	Input · $/1M	Output · $/1M
TokenKiln Default	tokenkiln/default	$0.50	$1.50

How it works

1
Top up
Buy prepaid credits with Stripe Checkout — balance is tracked to the micro-dollar.
2
Mint a key
Create scoped API keys with optional monthly spend limits. Keys are hashed at rest.
3
Call the API
Point any OpenAI SDK at the gateway. Every request settles against real engine usage.

One request away

curl https://your-gateway.up.railway.app/api/v1/chat/completions \
  -H "Authorization: Bearer tk_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tokenkiln/default",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": true
  }'

Works with the official OpenAI SDKs — just change baseURL and the API key.