Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

History

4 Commits

Repository files navigation

Aylin — AI Voice Agent Platform

Aylin is an open-source platform for building and deploying real-time AI voice agents. Users can create voice assistants by configuring instructions, LLM, TTS, and STT providers — no code required.

Built on LiveKit Agents, Aylin handles the infrastructure so you can focus on building your agent's behavior.

Architecture

aylin/
├── apps/
│ ├── backend/ # Node.js + Express REST API & LiveKit token server
│ ├── frontend/ # React + TypeScript dashboard UI
│ └── agent/ # Python LiveKit agent worker
└── azure-deploy/ # Kubernetes manifests & deployment scripts for AKS

The three apps run independently and communicate over HTTP/WebSocket:

Browser ──► Frontend (Vite, :5173 dev / :80 prod)
│
▼
Backend (:3000) ──► PostgreSQL
│ │
▼ │
LiveKit Cloud (agent reads
│ config from DB)
▼
Agent Worker (Python)

Features

  • Create, edit, and delete voice agents with a web UI
  • Configure per-agent instructions, voice, and model settings
  • TTS providers: OpenAI, Cartesia
  • STT providers: OpenAI, AssemblyAI
  • LLM providers: OpenAI
  • LiveKit real-time audio integration
  • Google OAuth login
  • Google Calendar and Sheets integration
  • Phone call support via LiveKit SIP including inbound and outbound calls
  • Telephony integration with Twilio and trunk providers like Telnyx
  • Azure Kubernetes deployment scripts included

Prerequisites

RequirementVersionNotes
Node.js18+For backend & frontend
Python3.9+For the agent worker
PostgreSQL14+Local or hosted
LiveKitAnyLiveKit Cloud (free tier available)

API keys you'll need:

  • LiveKit — API key & secret (required)
  • OpenAI — API key (required for LLM/TTS/STT)
  • Cartesia — API key (optional, for Cartesia TTS)
  • AssemblyAI — API key (optional, for AssemblyAI STT)
  • Google OAuth credentials (optional, for Google login)

Quick Start

1. Backend

cd apps/backend
npm install
cp .env.example .env
# Fill in your LiveKit credentials and database URL in .env
npm run dev

The backend starts at http://localhost:3000.

2. Frontend

cd apps/frontend
npm install
cp .env.example .env
# .env.example works as-is for local development
npm run dev

The frontend starts at http://localhost:5173.

3. Agent Worker

cd apps/agent
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env
# Fill in your LiveKit and AI provider API keys in .env
python agent.py dev

4. Open the app

Go to http://localhost:5173, sign in, and create your first agent from the Agents page. Test it live in the Playground.

Environment Variables

Backend (apps/backend/.env)

VariableRequiredDescription
LIVEKIT_URLYesLiveKit server WebSocket URL
LIVEKIT_API_KEYYesLiveKit API key
LIVEKIT_API_SECRETYesLiveKit API secret
DATABASE_URLYesPostgreSQL connection string
JWT_SECRETYesSecret for signing JWT tokens
SESSION_SECRETYesSecret for session cookies
GOOGLE_CLIENT_IDNoGoogle OAuth client ID
GOOGLE_CLIENT_SECRETNoGoogle OAuth client secret
LIVEKIT_SIP_ENDPOINTNoLiveKit SIP URI for phone call support

Agent (apps/agent/.env)

VariableRequiredDescription
LIVEKIT_URLYesLiveKit server WebSocket URL
LIVEKIT_API_KEYYesLiveKit API key
LIVEKIT_API_SECRETYesLiveKit API secret
OPENAI_API_KEYNoOpenAI API key
CARTESIA_API_KEYNoCartesia TTS API key
ASSEMBLYAI_API_KEYNoAssemblyAI STT API key

Frontend (apps/frontend/.env)

VariableRequiredDescription
VITE_API_URLYesBackend API base URL (default: http://localhost:3000/api)

Tech Stack

Backend: Node.js, Express, TypeScript, PostgreSQL, LiveKit Server SDK

Frontend: React 18, TypeScript, Vite, React Router, i18next, LiveKit Components

Agent: Python 3.9+, LiveKit Agents SDK, OpenAI, Cartesia, AssemblyAI

Deployment

Kubernetes manifests and deployment scripts for Azure AKS are in azure-deploy/.

A GitHub Actions workflow is included for CI/CD to Azure Container Registry + AKS.

Contributing

Contributions are welcome. Please open an issue to discuss what you'd like to change before submitting a pull request.

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/my-feature)
  3. Commit your changes
  4. Open a pull request

License

MIT

About

Aylin is an open-source platform for building and deploying real-time AI voice agents. Users can create voice assistants by configuring instructions, LLM, TTS, and STT providers — no code required.

Topics

Resources

Stars

2 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages