2FA fails at platform.openai.com

“Oops. Your authentication token is no longer valid”

This happens on Firefox with the Google OTP and with the Email OTP.

I have tried to disable any blockings from trackers to platform-openai-com and auth0-openai-com

Looking at networkanalysis everything looks fine but on step for 2FA

Loading of this here fails with NS_BINDING error
tcr9i-chat-openai-com/fc/a/?callback=__jsonp_xxx&category=loaded&action=game%20loaded&session_token=xxx&data[public_key]=xxx&data[site]=undefined

When I look at the browser requests of the page after the submit
POST auth0-openai-com/u/mfa-otp-challenge?state=hKFo2SB0Nxxx
GET auth0-openai-com/authorize/resume?state=tYIfKxxx
GET platform-openai-com/auth/callback?code=2kUIYPfykxxx&state=bC5xZXRmakxxx%3D%3D

In the console the following error shows up


{
	id: "embeddings",
	title: "Embeddings",
	content: "Get a vector representation of a given input that can be easily consumed by machine learning models and algorithms.\n\nRelated guide: [Embeddings](/docs/guides/embeddings)\n",
	url: "/docs/api-reference/embeddings",
	sections: [{
		type: "object",
		title: "The embedding object",
		content: "Represents an embedding vector returned by embedding endpoint.\n",
		url: "/docs/api-reference/embeddings/object",
		relativeUrl: "embeddings/object",
		definition: {
			type: "object",
			description: "Represents an embedding vector returned by embedding endpoint.\n",
			properties: {
				index: {
					type: "integer",
					description: "The index of the embedding in the list of embeddings."
				},
				embedding: {
					type: "array",
					description: "The embedding vector, which is a list of floats. The length of vector depends on the model as listed in the [embedding guide](/docs/guides/embeddings).\n",
					items: {type: "number"}
				},
				object: {
					type: "string",
					description: 'The object type, which is always "embedding".',
					enum: ["embedding"]
				}
			},
			required: ["index", "object", "embedding"],
			"x-oaiMeta": {
				name: "The embedding object",
				example: '{\n  "object": "embedding",\n  "embedding": [\n    0.0023064255,\n    -0.009327292,\n    .... (1536 floats total for ada-002)\n    -0.0028842222,\n  ],\n  "index": 0\n}\n'
			}
		}
	}, {
		type: "endpoint",
		title: "Create embeddings",
		url: "/docs/api-reference/embeddings/create",
		relativeUrl: "embeddings/create",
		method: "post",
		path: "/embeddings",
		definition: { /* left out */ }
	}]
}

Login works on Chrome

I have compared the browser requests on Chrome and they look the same

Also the code snipped in the main.js is the same. It seems that there is something wrong with this third value of embedding. Looks like some status message. Firefox fails with that and Chrome handles it different.

I have further analyzed this issue and have reformatted the main.js in PhpStorm
The problematic code then is around line 48580

It appears that the problem actually exists on the documentation of the api reference

Using ChatGPT to find a solution it actually found this ticket :wink:
Uploading the reformatted file it was able to come to this conclusion

The code segment from the main.js file around lines 48547 to 48727 appears to be related to the documentation or description of an “embeddings” feature. It mentions getting a vector representation of an input for machine learning models. It seems to be intended as a documentation example rather than executable code. It’s unusual for such a detailed example to be embedded directly in a JavaScript file, as it’s not standard practice and could potentially lead to confusion or errors during execution, especially in different browser environments.