rising-crimson•15mo ago
How To Send Voice Message to Voice Assistant
I have created the chatbot UI in my Flutter mobile app. I am using a Voice assistant, how do I send voice message to the Dialog Manager API so that it can continue with the conversation? The current API only accept text parameter which does not fit my use case. The voice feature works fine in Voiceflow website but I could not utilize it in my app, can anyone help?
I have used speech-to-text library but it can only record 5 seconds and no way to extend it to record longer question. I really have to send a long text question to ask the bot, the only way I can think of is to record as audio file and send to Voiceflow but I don't see it accepts voice input from the API.
3 Replies
Voiceflow provides a platform for designing, prototyping, and developing conversational assistants. These assistants can support two modes of communication: chat and voice. Once you've created an assistant, you can design the conversational flow and then share, test, and deploy the assistant.
You can integrate your Voiceflow assistant with various channels like WhatsApp and Discord, or even voice-based interfaces like Twilio IVR. To do this, you can leverage the Dialog Manager API. This API allows you to send input to your Voiceflow agent and fetch a response based on your design. It can also manage the state of your user's conversation, allowing them to leave the conversation and then pick it up later from the same spot.
If you're looking to customize the chat widget experience, you can use Voiceflow's open-source Web Chat project called react-chat. Note that a custom launch solution requires the ability to host and maintain code on your own infrastructure.
For more detailed information, you can refer to the Voiceflow Dialog Manager API documentation.
rising-crimsonOP•15mo ago
how do I send voice message through Dialog Manager API so that voiceflow can do the translation?
foreign-sapphire•14mo ago
There is no STT endpoint in voiceflow. You'll have to use Whisper or any other STT service.
Here's a demo of the architecture: https://github.com/voiceflow-gallagan/asr-demo-whisper
Unfortunately, there is no documentation on how to add voice capabilities to the Web Chat SDK. Maybe @Voiceflow Connect or @Tico | Voiceflow Assistant can help with some tips. I'm going to need to do this soon for a project so I'd appreciate your help.