Release ChatGPT usage budgets hourly, daily, or weekly

This runnable example automates ChatGPT Enterprise and Edu usage-limit updates with the ChatGPT Admin API. It includes a Node.js program you can try locally, then configure for your workspace.

Usage limits in ChatGPT Enterprise and Edu are monthly. A user who spends their allocation in the first week reaches the limit with most of the month still ahead.

Release portions of a budget in credits or US dollars (USD) hourly, daily, weekly, or on a custom schedule to keep capacity available for later work. Choose amounts and schedules for individual users, a department, or your workspace.

For your most active users, increase the available amount based on recent consumption, up to a monthly maximum you set. Pair these adjustments with regular usage updates so people understand what remains, when more will be available, and how their model choices affect consumption.

Keep capacity available throughout the month

The illustration uses 2,000 credits per person per month. All amounts in this guide are examples; choose budgets that fit the people and teams you support.

Releasing 500 credits each week reserves part of that allocation for later weeks. Each release increases the person’s monthly limit, reaching the full 2,000 credits with the fourth release.

Monthly reset dates

Usage limits reset on the first day of each month in UTC by default. A workspace can also align its monthly usage period with its billing cycle. See your workspace’s usage period.

Try a release plan in your browser

Open the interactive example. Choose Credits, US dollars (USD), or Both to compare two separate example plans. Change the budgets and release schedules, and see how usage affects the amount available.

Release date	New credits available	Monthly limit after release
Day 1	500	500
Day 8	500	1,000
Day 15	500	1,500
Day 22	500	2,000

Unused released credits remain available within the month. You choose the total monthly budget and when each portion becomes available.

Choose dollars or credits

Use the billing unit shown in your workspace’s usage settings for your customer agreement. Choose Credits for a credit-based workspace or US dollars (USD) for a workspace with dollar-denominated usage and limits. The controller checks the API’s reported unit before applying a plan.

Example plan	Monthly budget per person	Weekly release	Configuration
Credits	2,000 credits	500 credits	"unit": "credit"
US dollars	$200.00	$50.00	"unit": "usd"

These are separate illustrative budgets. Choose amounts for your organization in its billing unit. Credit limits use whole credits; dollar limits use cents. All three release approaches support both units. Configure the unit and amounts.

Get started

You need Node.js 24 or later to run the local example on Windows, macOS, or Linux. It uses three fictional people and keeps all changes in memory. You do not need an Admin API key.

1. Download the starter ZIP and extract it.
2. Open the chatgpt-usage-budget-starter folder. It contains README.md, docs, src, and assets.
3. Open a terminal in that folder. On Windows, right-click the folder’s background in File Explorer and choose Open in Terminal. On macOS or Linux, open Terminal, type cd , drag the extracted folder into the window, and press Enter.

Run the credit example:

node src/demo.mjs --unit credit

Or run the dollar example:

node src/demo.mjs --unit usd

The program previews the first release, applies it in the local simulator, and advances to the next weekly release. Each fictional person’s limits follow the selected plan:

Example	Starting monthly limit	First release	Next weekly release
Credits	2,000 credits	500 credits	1,000 credits
US dollars	$200.00	$50.00	$100.00

The demonstration also simulates an interrupted update and a conflicting admin edit, then restores the starting settings. A successful run ends with Demo complete. All simulated state stayed in memory.

Choose the setup path that fits the tools available to you:

Your setup	Guide
You use Codex and want help with setup.	Open the folder in Codex and ask it to run the example.
Your team provides a managed computer, virtual machine, or cloud environment.	Run in that environment. This path also works when you cannot run the required software on your computer.

After the demonstration, choose the people, amounts, and schedule for your pilot.

How the weekly plan uses the API

This configuration excerpt expresses the 500-credit weekly plan. Include these fields in the complete configuration created by the configuration and enrollment guide, and replace the example anchor with the start of your plan within the confirmed monthly period.

{
  "unit": "credit",
  "policy": {
    "pattern": "fixed_release",
    "anchor": "2030-01-01T00:00:00Z",
    "intervalHours": 168,
    "startCap": "500",
    "increment": "500",
    "ceiling": "2000"
  }
}

startCap sets the first monthly limit. Every 168 hours, the controller adds increment to the scheduled target, up to ceiling. The resulting monthly limits are 500, 1,000, 1,500, and 2,000 credits. A repeated run in the same week keeps the same target; a missed run catches up to the current week’s target.

For the second release, the API adapter sends a request with this shape after the enrollment and proposed changes have been reviewed. WORKSPACE_ID and USER_ID below stand for the selected workspace and enrolled person:

PATCH /v1/manage/workspaces/WORKSPACE_ID/usage_limits/users/USER_ID
Host: api.chatgpt.com
Content-Type: application/json

{
  "override_monthly_usage_limit": {
    "type": "limited",
    "limit_amount": { "unit": "credit", "amount": "1000" },
    "temporary": true
  }
}

The request sets that person’s total monthly limit to 1,000 credits. The workspace balance and monthly usage period stay unchanged. temporary: true makes the override expire at the end of that period. Continue with the reviewed enrollment and preview steps before applying a plan to your workspace; the API contract covers authentication, units, and restoration.

Choose how to release the budget

Approach	Use it to
Scheduled releases	Spread a monthly allocation across hourly, daily, weekly, or custom releases.
Usage-based increases	Give active users more capacity as consumption grows, up to a monthly maximum.

Set different budgets and schedules for different populations. For example, give a project team weekly releases and use consumption-based increases for a small group of power users. Keep each person in one active configuration.

See policy configuration and examples for the settings. Share the monthly allocation, usage so far, and next release date through your existing communication channels. Use those updates to help people choose models that fit their work and available budget.

Test with a small group

Use a local run or Codex automation to try the approach with a small group. Review the proposed limits, observe how the release schedule fits their work, and adjust the plan.

For ongoing workspace budget management, run the program on organization-managed infrastructure, such as a managed virtual machine or cloud service. Assign a team to monitor it and keep it running independently of an individual’s computer.

Next step	Guide
Test a scheduled Codex pilot	Set up a Codex pilot
Run on a managed macOS or Linux computer or server	Set up a managed host
Deploy using your AWS account	Deploy on AWS

The AWS guide includes a deployment template, queued processing for larger enrollments, and progress tracking. Start with a small group, then increase the enrollment as you verify completion and API throughput. Your technical team can adapt the program to other infrastructure your organization operates.

For Windows admins, the AWS guide provides a browser-based setup path through AWS CloudShell. Live local scheduling uses the included macOS or Linux service setup.

Prepare a reviewed enrollment

Choose a workspace, then select people by user ID, email address, workspace group ID, or all current members. Review the resolved list, current settings, monthly period, and proposed changes. The configuration and enrollment guide takes you through each step. A preview shows the proposed limits before you apply them.

Stop, restore, and renew

The program saves the original settings and records its changes. Follow the stop and restore procedure when ending a pilot or changing its policy. For the next monthly period on AWS, confirm its dates and review the opening limits. The renewal helper carries forward your policy and people on the same deployment, with its existing records and credentials.

For implementation details, see the verification guide and API contract.