Assistant bugging out
# support
l
My assistant got stuck in a loop and started acting sporadically. I cant have this happening during our demo next week. call id 1: a2b1bd9b-b305-4584-9937-35857196eac0 call id 2: 0fe38c34-c3e0-47e2-9cef-fdafabc5b3dc - didnt respond for 20 seconds then got stuck in a loop then started reading out the exact prompt including asterix's Could I please have some insight into the issue Thanks
v
Hi @Louis, I'm an AI assistant for Vapi. I'll analyze available sources and get back to you shortly.
----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Introduction]() * [List Calls]() * [Get Call]() * [Web Calling]() -----------------------------
### Answer It sounds like you are experiencing issues with your assistant due to potential misconfigurations or model limitations. Here are a few suggestions to troubleshoot and improve the reliability of your assistant: 1. **Limit Response Length**: Ensure your system prompt includes a directive to keep responses concise. For example, you can add:
Copy code
PLEASE DON'T RESPOND IN MORE THAN 2-3 SENTENCES. THIS IS A VOICE CONVERSATION.
This can help prevent the assistant from getting stuck in long responses [(1)]() . 2. **Use a Reliable LLM**: Consider switching to a more reliable model like GPT-4 if you're currently using a lower-tier model. Lower-tier models may hallucinate and behave unpredictably [(2)]() . 3. **Check Configuration**: Ensure your assistant's configuration is set up correctly. You can create an assistant configuration object and pass it to the
.start()
method:
Copy code
javascript
    const assistantOptions = {
        name: "Demo Assistant",
        firstMessage: "Hello, how can I assist you today?",
        transcriber: {
            provider: "deepgram",
            model: "nova-2",
            language: "en-US",
        },
        voice: {
            provider: "playht",
            voiceId: "jennifer",
        },
        model: {
            provider: "openai",
            model: "gpt-4",
            messages: [
                {
                    role: "system",
                    content: `You are a voice assistant for a demo. Please keep responses short and to the point.`,
                },
            ],
        },
    };

    vapi.start(assistantOptions);
[(3)]() . 4. **Function Calls**: Ensure that function calls are returning responses promptly. If a function does not return a response, the assistant may hang or loop [(4)]() . By following these steps, you can potentially mitigate the issues you're facing with your assistant during the demo. SOURCES USED: [(3)]() , [(1)]() , [(2)]() , [(4)]()
s
Looking into it.
2 Views