Create Simli App (Groq + Deepgram)

This starter is an example of how to create a composable Simli interaction that runs in a Next.js app. The project consists of a Next.js app that uses the Simli SDK (simli-client) and a server server.js that handles the interaction with other services such as speech-to-text (STT), large language models (LLMs) and text-to-speech (TTS).

Environment variables

Start by signing up and getting your API key from Simli.com. Then, fill in the .env file in the root of the project and put in the following environment variables:

NEXT_PUBLIC_SIMLI_API_KEY="API key from simli.com"
DEEPGRAM_API_KEY="API key from deepgram.com"
GROQ_API_KEY="API key from console.groq.com"

If you want to try Simli but don't have API access to these third parties, ask in Discord and we can help you out with that (Discord Link).

To run the back-end and front-end together, run the following command:

npm run start

Characters

You can swap out the character by finding one that you like in the docs, or create your own

Alternative STT, TTS and LLM providers

You can of course replace Deepgram with AI services with your own preference, or even build your own. The only requirement for Simli to work is that audio is sent using PCM16 format and 16KHz sample rate or sending it through MediaStream. If you're having trouble getting nice audio, feel free to ask for help in Discord.

Links

[Simli]
[Deepgram] [Groq]

Deploy on Vercel

An easy way to deploy your avatar interaction to use the Vercel Platform.