How to Manage Serverless Inference Prepayment

Last verified 1 Oct 2026

Inference provides a single control plane for managing inference workflows. It includes a Model Catalog where you can view available foundation models, including both DigitalOcean-hosted and third-party commercial models, compare model capabilities and pricing, use routing to match inference requests to the best-fit model, and run inference using serverless or dedicated deployments.

Serverless Inference requires promotional credits or a positive prepaid balance before you can send inference requests. You can fund it with either of two balances on your team:

  • Inference and Agents balance: Only Serverless Inference and Managed Agents draw from this balance, so charges from other DigitalOcean products never reduce it. Use it to keep a separate budget for AI workloads.
  • Account prepayment balance: This balance applies to your entire account, so charges from other DigitalOcean products also draw from it.

DigitalOcean deducts usage charges as you incur them. When your promotional credits and both balances are empty, DigitalOcean suspends your access to Serverless Inference until you add funds. For the order in which usage draws from your credits and balances, see How Your Prepaid Balances Are Shared.

You can view both balances on the Billing page. To keep inference running, add funds to your Inference and Agents balance or add an account prepayment, and enable auto-reload to replenish your balance automatically.

Add Funds to Your Inference and Agents Balance

On the Billing page, in the Inference & Agents balance section, click Add funds, then choose a preset amount from $5 to $500. Auto-reload is on by default on that page, so turn off Enable auto-reload if you want a one-time payment. For details, see Add Funds to the Inference and Agents Balance.

After your payment processes, the balance updates and Serverless Inference access resumes if it was suspended.

Add an Account Prepayment

A prepayment is a one-time payment that adds funds to your account balance ahead of your usage. Charges from all DigitalOcean products draw from this balance. To add a prepayment, use the Make a payment page, where you can enter an amount and pay with a saved payment method or a third-party provider.

After your payment processes, your account prepayment balance updates and Serverless Inference access resumes if it was suspended.

How Your Prepaid Balances Are Shared

Serverless Inference, Harness Runtime, and the Action Gateway tools that require prepayment can draw from two prepaid balances on your team:

  • The Inference and Agents balance, which only Serverless Inference and Managed Agents draw from. Charges from other products never reduce it.
  • The account prepayment balance, which applies to your entire account (your current team). Usage charges from all DigitalOcean products, such as Droplets or Managed Databases, draw from it.

Usage draws from your credits and balances in this order:

  1. Promotional credits.
  2. The Inference and Agents balance.
  3. The account prepayment balance.

Promotional credits apply to Serverless Inference, Harness Runtime, and Action Gateway, so these products keep running while you have credits, even if both prepaid balances are at $0. When your credits and both prepaid balances are empty, DigitalOcean stops prepaid spend for these products until you add funds. Action Gateway tools that do not require prepayment are not blocked by the gateway’s balance check.

Other products continue running as usual, even if their charges are what depleted the account prepayment balance. Those products keep accruing usage against your account.

For example, if you add a $500 prepayment to your account:

  1. You run Serverless Inference workloads that cost $100, leaving a balance of $400.
  2. Your Droplets accrue $250 in usage charges, leaving a balance of $150.

Only the remaining $150 is available for prepaid products. If charges from other products bring the account prepayment balance down to $0, DigitalOcean stops prepaid spend for those products until you add more funds, but your Droplets and other resources continue running without interruption.

If you add the same $500 to your Inference and Agents balance instead, the $250 in Droplet charges bill on your monthly invoice, and $400 remains available for prepaid products.

To keep a separate prepaid budget for these workloads, add funds to your Inference and Agents balance. To avoid interruptions, enable auto-reload to replenish the balance automatically.

You can also create a dedicated team for these workloads. Each team maintains its own prepaid balances, so charges from resources on your other teams do not draw them down.

Note

Teams that belong to an organization draw from the organization’s Inference and Agents balance and account prepayment balance, which the organization funds. They do not support team-specific balances. To keep a separate balance, create a standalone team outside of your organization.

Enable Auto-Reload

Auto-reload keeps Serverless Inference running by adding funds to your Inference and Agents balance when it reaches a threshold you set. To turn it on, on the Billing page, in the Inference & Agents balance section, click Turn on auto-reload. For the full steps, see Enable Auto-Reload.

If you already use auto-reload on your account prepayment balance, it keeps running until you turn it off on the Billing page.

We can't find any results for your search.

Try using different keywords or simplifying your search terms.