Pricing

Pay for what you use. Nothing else.

Self-host for free, or let us run it with your own provider keys for one flat 3% fee on usage, billed monthly. No fixed fee, no seats, nothing to prepay.

Self-hosted

Run the open-source gateway on your own infrastructure.

FreeMIT license, forever
  • Every gateway feature
  • Your servers, your data
  • Docker and Helm installs
  • Community support on GitHub
Bring your own keys

Cloud

The same gateway, hosted and run for you.

3%of usage cost, billed monthly
  • Bring your own provider keys
  • Your providers bill you for tokens directly
  • API keys, budgets and usage dashboard
  • No fixed monthly fee, no seat fees
  • Nothing to prepay, cancel any time
  • Email support

Enterprise

For teams with volume, compliance or deployment needs.

Customtalk to us
  • Volume discounts on the platform fee
  • Pay by invoice
  • Priority support
  • Help running DeltaLLM in your cloud

Estimate

See what you'd pay

Move the slider to your expected monthly usage cost. The fee is the same 3% at every size.

Monthly usage cost

$100

$10$10,000

The usage cost of your requests at each provider's public list price. Your providers bill you for the tokens directly.

Usage cost at provider list prices
$100
DeltaLLM platform fee (3%)
$3.00
Charged by DeltaLLM, after the month$3.00

Estimate before any sales tax or VAT. Paid to your providers separately, under your own accounts.

Usage cost

How we calculate the 3%

The fee is based on what your requests would cost at each provider's public prices, so it is the same whatever deal you have with your providers.

1. We count what each request used

For every request that goes through the gateway, we record the tokens the provider reports: input, output, and other billed units such as cached tokens where the provider reports them.

2. We price it at the provider's list price

The usage cost is those tokens multiplied by the provider's public list price for that model on the day of the request. If we have no list price for a model, we use a default rate of $1 per million input tokens and $4 per million output tokens. The console shows the rate for each of your models before you use it.

3. We take 3% of the month's total

At the end of each monthly billing period, we add up the usage cost of all your requests and charge 3% of it. You can see the running total in the console at any time.

Example: 10 million input tokens at $3 per million and 2 million output tokens at $15 per million is a usage cost of $60, so the fee is $1.80.

Cloud

Every charge, in one table

Platform fee

Charged to your saved payment method once a month, for the previous month's usage.

3% of usage cost

Model tokens

Your AI providers bill you directly under your own accounts. DeltaLLM does not sell model usage.

Paid to your provider

Sales tax or VAT

Added by our payment provider based on where you are, and shown on your receipt.

Where it applies

Fixed monthly fee, seats, setup, prepaid balance

A month with no usage costs nothing.

None

Compare

Self-hosted, Cloud or Enterprise

FeatureSelf-hostedCloudEnterprise
Gateway
OpenAI-compatible API
OpenAI, Anthropic, Google, Groq, Azure and more
Routing, fallbacks and load balancing
Bring your own provider keys
Control
API keys per app or teammate
Budgets and spend limits
Usage and cost dashboard
Sign in with Google
Operations
Hosting, upgrades and monitoringYouUsUs or you
SupportCommunityEmailPriority
Billing
Platform feeNone3%Volume discount
Fixed monthly feeNoneNoneCustom
PaymentNoneCard, monthly after usageCard or invoice

FAQ

Billing questions

What is the 3% calculated on?
On the usage cost of your requests at each provider's public list price, not on your provider invoice. If you have discounts, free tiers or negotiated rates with a provider, your provider bill can be lower than the usage cost we use, and the fee stays 3% of the list-price cost. Requests that the provider rejects or that return no tokens add nothing.
When am I charged?
Your monthly billing period starts on the day you add a payment method. At the end of each period, we charge the fee for that period's usage to your saved payment method and email you a receipt. If the fee is zero, we charge nothing. If it is under $1, we add it to the next period's charge instead of charging it on its own.
What am I agreeing to when I add a payment method?
You authorize us, through our payment provider, to charge that payment method once a month, after each billing period, for 3% of that period's usage cost plus any tax. The amount changes with your usage. You can set spend limits on your API keys, and you can cancel at any time.
How do I cancel?
Cancel from the Billing page in the console, or email contact@deltallm.io. Your API keys stop working right away, we charge one last time for the fee on usage up to the moment you cancelled (we waive it if it is under $1), and there are no charges after that. There is no cancellation fee.
What if I'm billed incorrectly?
If we charge you more than these terms allow, for example a duplicate charge, a wrong price, a charge after you cancelled, or usage from requests that failed because of a fault on our side, we refund the difference to your original payment method. See the refund policy.
What does the fee cover?
Running the gateway for you: hosting, routing and fallbacks, API keys, budgets, usage tracking and support.
Is self-hosting really free?
Yes. The gateway is open source under the MIT license, with no usage limits. You pay only for your own servers and provider usage. Get it on GitHub.

Connect your keys and start

Sign in with Google and send your first request in minutes.