Skip to main content

​
After this article

You will be able to connect OpenAI to Lumos and sync users, projects, and role data.

​
Capabilities

​
Required plan and roles

  • OpenAI role: you need Owner or Admin permissions in the organization to create an Admin API key.
  • Admin API key: the key must be an organization-level Admin API key, not a personal or project key. Admin keys are used for programmatic administration of your OpenAI account and grant access to the endpoints in the API Reference for Organizations.

​
Read-only access vs provisioning

Is read-only possible? Yes, and it’s the only mode. This connector only syncs users, project roles, and roles. What read-only requires. An organization-level Admin API key (not a personal or project key) created by a user with Owner or Admin permissions, plus your Organization ID, entered as described below. Watch out for. A personal or project-scoped key will authenticate but won’t return organization-wide user data, so confirm the key was created on the Admin keys page under an Owner or Admin account. The integration doesn’t track last login or activity, since OpenAI’s API doesn’t expose it.

​
Instructions

Phase 1: Create an Admin API key
  1. Sign into OpenAI as a user with Owner or Admin permissions.
  2. Go to your Admin keys page.
  3. Create an Admin API key.
  4. Copy the key so you can paste it into Lumos.
Phase 2: Find your Organization ID
  1. Open your Organization settings page.
  2. Copy your Organization ID.
Phase 3: Connect OpenAI in Lumos
  1. Click the OpenAI card in your Lumos integrations (reconnect or add new).
  2. Enter your Admin API Key.
  3. Enter your Organization ID.
  4. Click Connect.

​
API Endpoints Used

​
Troubleshooting

Users are not syncing. OpenAI organization members are not appearing in Lumos. Verify that your API key is an organization-level key (not a personal or project key). Check that the key has not been revoked or deleted. Ensure you have Owner or Admin permissions in the organization. Confirm members have accepted their organization invitations. Review the Lumos integration logs for specific error messages. API key authentication fails. If you see “Invalid API key” or “Authentication failed” errors, verify you copied the complete API key without extra spaces, ensure the key is for the correct organization, and check that it has not been revoked in OpenAI. Confirm the key is an organization key, not a project or personal key. Generate a new key if the original has been compromised, and verify your OpenAI account is active and not suspended. Organization not found. If you see “Organization not found” or “Invalid organization” errors, verify you are using an organization-level API key and confirm you are part of an OpenAI organization (not just a personal account). Check that the Organization ID is correct, and ensure the correct organization was selected when the key was created. Contact OpenAI support if organization access issues persist. Some user details are missing. Verify that users have complete profiles in OpenAI and have confirmed their email addresses. Ensure users have accepted their organization invitations, and confirm the API key has read permissions for user details. Rate limit errors. If the integration fails with “Rate limit exceeded” errors, wait for the rate limit window to reset (typically 1 minute to 1 hour). Check whether multiple integrations use the same API key, review your OpenAI API usage dashboard, and consider upgrading your OpenAI plan if you consistently hit limits. Contact OpenAI support to increase rate limits if needed. The connector is blocked by a network proxy. Lumos cloud connectors egress from the Lumos IPs 44.228.74.159 and 54.244.99.235. Allowlist both (for example in Zscaler).