Every call, priced
See per-call cost by model, product, customer and tag. The total is made of rows you can inspect.
gpt-4.1-mini
search · customer_042
gemini-2.5-pro
support · #production
THE AI GATEWAY WITH THE RECEIPTS
One gateway for OpenRouter, TrustedRouter, OpenAI and Gemini. Every call logged, tagged and priced, with your own keys.
YOUR KEYS. YOUR MODELS. THE NUMBERS IN ONE PLACE.
EVERY PROMPT HAS A PRICE.
/01 SETUP
Point your OpenAI client at Untokened. Keep your existing calls; get a receipt for each one.
Read the docsfrom openai import OpenAI
client = OpenAI(
api_key="YOUR_UNTOKENED_KEY",
base_url="https://api.untokened.com/v1",
)
response = client.chat.completions.create(
model="gpt-4.1-mini",
messages=[{"role": "user", "content": "Hello"}],
)/02 PRODUCT
See per-call cost by model, product, customer and tag. The total is made of rows you can inspect.
gpt-4.1-mini
search · customer_042
gemini-2.5-pro
support · #production
Connect OpenRouter, TrustedRouter, OpenAI and Gemini. Keys stay in your vault, and you can switch upstream per product.
UPSTREAM / CONNECTED
Fetch a prompt by name and label. Compare cost and latency per version before you move the label.
VERSION
v12
AVG COST
$0.0084
LATENCY
842 ms
Group multi-step agent runs into one trace. Follow each call from first question to final answer.
Set limits per key, product or customer. Get alerts as spend approaches the line, and stop calls when it crosses it.
Store data in EU regions. Route regulated traffic to EU and zero-retention providers when the request requires it.
/03 HOW IT WORKS
01 / SOURCE
Your apps
Products · agents · services
02 / ROUTE
Untokened gateway
Log · tag · price
03 / UPSTREAM
Model providers
OpenRouter · TrustedRouter · OpenAI · Gemini
Your dashboard
/04 NEXT STEP
Every call has a cost. See it before the month ends.