setwitch

Hello,

Thank you for your first reply, which helped us understand that the Quick Test does not generate any token usage, since it simply checks that the API key works and retrieves the list of available models.

However, our main issue remains unresolved.

We created a WordPress test page with our AI Engine chatbot and made actual requests through the chatbot.

The behavior is as follows:

• the OpenAI connection is validated by Quick Test: Connection successful;
• AI Engine shows that the OpenAI environment is configured;
• our chatbot is configured to use this environment;
• however, when we send an actual request to the chatbot, the exact response displayed is:

OpenAI (Responses API): You have no credits remaining. Add credits to continue using the API at https://platform.openai.com/settings/organization/billing/.

We have attached screenshots of the chatbot showing this response.

We therefore have several important questions.1. What exactly does this “credit” refer to?

Does this error message concern:

• the OpenAI API account balance/credits;
• credits from another provider;
• a limit related to the OpenAI project;
• a limit related to the API key;
• a limit or configuration specific to AI Engine;
• or possibly the way our chatbot is configured, with a homepage that allows the user to select a language?

We want to be certain that we understand which account or limit is actually involved.2. Does the chatbot request actually reach OpenAI?

Since the Quick Test works, we know that the API key is recognized.

However, we would like to verify that an actual request generated by the AI Engine chatbot is indeed being sent to OpenAI and that it is using the same environment and the same API key.3. Is our AI Engine configuration correct?

We are attaching screenshots of our AI Engine configuration, including:

• Dashboard
• Modules
• Chatbots
• Settings → AI
• Settings → Chatbot
• Settings → MCP
• Settings → Files & Media
• Settings → REST API
• Settings → Others
• Development Tools
• License

We are also attaching a screenshot showing the Connection successful result from the OpenAI connection test.4. We also have usage visible in OpenAI

In our OpenAI account, we noticed a request showing approximately 428 input tokens.

We would like to know whether this usage could correspond to a request made by our AI Engine chatbot, or whether it necessarily came from another use of the API key.

If necessary, we can create a dedicated API key exclusively for AI Engine so that we can identify its requests precisely.5. Additional information

If necessary, we can also provide:

• any AI Engine logs;
• any request ID that may be available;
• any other screenshot or information that would help you analyze our problem precisely.

Our goal is simply to understand why an actual chatbot request is being rejected while the Quick Test using the same OpenAI configuration is successful.

Please let us know precisely where the problem lies and which credit or limit the chatbot error message is referring to.

Thank you very much for your help.

SCREEN CAPTURES

WP - AI Engine - Chatbot
https://setwitch.com/wp-content/uploads/2026/10/WP-AI-Engine-Chatbot.jpg

AI Engine - Dashboard
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Dashboard.jpg

AI Engine - Modules
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Modules.jpg

AI Engine - Chatbots
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Chatbots-scaled.jpg

AI Engine - Settings - AI
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Settings-AI.jpg

AI Engine - Settings - (Chatbot, MCT, Files & Media)
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Settings-Chatbot-MCT-Files-Media.jpg

AI Engine - Settings - (Rest API, Others)
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Settings-Rest-API-Others-scaled.jpg

AI Engine - (Development Tools, License)
https://setwitch.com/wp-content/uploads/2026/10/AI-Engine-Development-Tools-License.jpg

Open AI
https://setwitch.com/wp-content/uploads/2026/10/Open-AI.jpg