When it comes to voice AI agents, there are many TTS providers out there. But if you are trying to build real time speech solution and if you wish to build a service bot or AI agent with voice capabilities it's expensive.
However there are some of the providers like ElevenLabs which offers some credits to build the apps on their ecosystem.
I created a video explaining how to Create Voice AI agent using ElevenLabs API.
Here's how I start for any new voice AI agent.
Lets say you want to build voice agent for your store. First come up with data on your store. So that agent can pick the inventory, script and the data related to store.
Use langchain or Google ADK. And this framework would build a basic agent. You can use the Vercel Eve framework to build a chat agent around this product. And you can use LLM which gives decent processing for the text to voice conversion on the go.
This is the frame of your solution. You can extend this into the prompt and build the app. App is how your get agents to work. Building agent doesn't mean just writing the code on Google collab and be at that stage.
Build the implementation with an app. That's how you learn to build Voice AI agent.
Top comments (0)