Internal Error on First Login Only When Using Company Tenant (Works on Personal Tenant)

[biswajyotisaha]

Hi everyone,

I’m using Auth0 as the authentication provider for an application built with the ChatGPT Apps SDK, and I’m running into an issue that only occurs when using my company’s Auth0 tenant. The exact same setup works perfectly with my personal Auth0 tenant.

Issue Overview

When initiating the OAuth connection:

1. User is redirected to Auth0

2. User enters username/password

3. After clicking “Log In,” Auth0 redirects back to ChatGPT — but instead of completing the flow, the browser shows:

chatgpt.com/connector_platform_oauth_redirect?error=server_error&error_description=Internal+error&state=oauth_s_691c9a9817c88191a280e93965e37dd2

Important behavior notes:

• This error occurs only on the first login attempt for a user.

• If the same user immediately tries to log in a second time, the login works.

• If the user selects Sign Up instead of Sign In, the error does not occur at all.

• Using my personal Auth0 tenant, everything works (both first login and subsequent logins).

This makes the issue appear tenant-specific and timing-specific.

Auth0 Log Entry for the Failing Login

Auth0 logs show a generic Internal error without a clear cause. Here is the raw log entry from the company tenant (sensitive content removed):

{
  "date": "2025-11-18T16:11:39.684Z",
  "type": "f",
  "description": "Internal error",
  "connection": "Username-Password-Authentication",
  "client_id": "grx9sX6J6Gxu7kJNMvaduMV7GFnpao49",
  "client_name": "ChatGPT",
  "ip": "167.103.88.187",
  "user_agent": "Chrome 142.0.0 / Mac OS X 10.15.7",
  "details": {
    "error": {
      "message": "Internal error",
      "oauthError": "Internal error",
      "type": "server_error"
    },
    "session_id": "etAMe1j047Z2iyAGHcdsBEUFRCMwY02D",
    "riskAssessment": {
      "confidence": "high",
      "version": "1"
    },
    "actions": {
      "executions": [
        "ZTIFtPlmhc5Kl6aeGEo5DWAKMjAyNTExMThGmky8gl9DfY--ltxEPEO8"
      ]
    }
  },
  "hostname": "XXXX_XXXXX",
  "user_id": "auth0|6916cb85ab8fd0701235ddff",
  "strategy": "auth0",
  "strategy_type": "database",
  "audience": "XXXX_XXXXX",
  "scope": ["openid", "profile"],
  "tenant_name": "XXXX_XXXXX"
}

Unfortunately, this log doesn’t indicate why the internal error is occurring or what’s different about the first login.

Summary of the Problem

• Only the first login triggers server_error

• Second login works

• Sign-up works

• Works flawlessly on personal Auth0 tenant

• Company tenant logs only show a generic “Internal error”

What I’m Hoping to Learn

Has anyone encountered a scenario where:

• An Auth0 tenant throws an internal error only on the first login,

• But not on subsequent logins or sign-ups?

Could tenant-wide configurations — such as Actions, Rules, Attack Protection settings, first-login flows, user profile initialization logic, or organization membership checks — cause this behavior?

Any suggestions, similar cases, or debugging tips would be greatly appreciated.

Thanks in advance!

[biswajyotisaha]

OpenAI has changed their connector, Now we can add Client Id and Client Secret. So, now we can create a first party application in Auth0 tenant and use it for the connector. This prevents multiple dynamically created third party applications in the tenant. It also resolved my previously mentioned login issues.

[gerald.czifra]

Hi @biswajyotisaha

Thank you for sharing this update with the Community, we’re glad to know this is now resolved for you!

Have a great one!
Gerald