why in ai agent the voice i set and the voice coming is totally different and that voice is not even in option a creepy voice

Try this https://cdn.discordapp.com/attachments/1373022618739019958/1373023624293908490/image.png?ex=6828e765&is=682795e5&hm=38e170816a267da76157a58902c17a472628099f8e68b38bba5c3b4dd185d24b&

thhanks a lot bro

Make asistant and give the first message then he will start conversation.

bro i am using vapi as a model there i no option for first message

@ROSHAN https://cdn.discordapp.com/attachments/1373022618739019958/1373180294185877556/image.png?ex=6829794e&is=682827ce&hm=23daa6a02faec2aa1d50887825defa71e4ec70ab6ad816a62cfbedcd44f7e864&

bro but i am using vapi as provider

Means you are making work Flow

the say option is removed not there thats why the problem happening

just use conversatation node and select exact option and then give the first message

conversatation node have no exact option bro

This is a prompt cutomize it according to your need and give it in first node. You are a friendly assistant working at a pizza restaurant. Greet the caller and ask if they’d like to place an order or have a question. "Hi! Thanks for calling Vapi’s Pizzeria. Would you like to place an order or ask something?"

i want to store those values and then use it in api request but this is what i did but my api unable to fetch and understand and i am unable to create an required output https://cdn.discordapp.com/attachments/1373022618739019958/1373286600439758878/Screenshot_2025-05-17_183930.png?ex=6829dc4f&is=68288acf&hm=1171edfed84931b92ff20f0c106ef52b225709c4f2af2ae9b078630a20cb578f&

So basicaly now you want to store the answers of the interview and where are you storing the answers ...

i want to temporary store as we use to do in gather function

so you can use n8n for integrations

How do we get non-garbled, non -TTS-back-to-STT text back from the model as an API response? The messages returned in JSON by the API are garbled. However, Dashboard >

Call Logs

> select a call >

Logs

tab are great! Clearly pre-TTS straight from the model. Attached screenshots comparing the JSON response to our

Logs

tab https://cdn.discordapp.com/attachments/1373022618739019958/1373382376276889660/vapi-dashboard-log.png?ex=682a3582&is=6828e402&hm=aa5c6093a5bd05e8036765c73b28905279bd4a2d659238fdff7894efdab2438c& https://cdn.discordapp.com/attachments/1373022618739019958/1373382376616497333/vapi-json-response.png?ex=682a3582&is=6828e402&hm=3058ad1cb5c2563e3ab175daa5dd7af4d59cfa46f4f096b3d1879169268c7063&