I’m building a project that needs to access a custom external API that I’ve built.
I’ve tested the API and it’s returned some values.
But every time the Chatbot accesses the API, it’s never done processing a speech. It says “Response taking long to load. Please hold on tight” and be like that forever. Never returned a speech.
My API has confirmed receiving the request from Convai (as in the photo below)
The chatbot just won’t process its speech.
I assumed that because the JSON output is too long, but I guess that’s just not it. As you can see, I’ve tried reduced the fields of data down to only 2 (english_name & number). Still won’t work.
I’m using Claude and ChatGPT. Either won’t work.
Pls help, because this is crucial to my project. It needs accurate data.
Could you try adding "additionalProperties": false at line 14 in your input description. And testing again with the GPT-4o model.
The final input_description could look something like following:
{
"parameters": {
"lines": {
"type": "array",
"description": "An array of 6 objects representing the lines of the hexagram. Each object must include position (1-6), value ('yin' or 'yang'), and changing (true or false).",
"items": {
"type": "object",
"properties": {
"position": { "type": "integer", "description": "Line position from bottom (1) to top (6)" },
"value": { "type": "string", "description": "'yin' (broken) or 'yang' (solid)" },
"changing": { "type": "boolean", "description": "Whether this line is changing" }
},
"required": ["position", "value", "changing"],
"additionalProperties": false
}
}
},
"required": ["lines"]
}
Please test it again using the GPT-4o model and let me know if that resolves the issue for OpenAI. If it still hangs or fails to generate a speech response, especially when using your external API, feel free to share any error logs or specific behaviour you’re noticing — that’ll help us debug further.
