Meow-AI is an intelligent real-time voice companion designed to provide empathetic listening and emotional support. It creates a safe, judgment-free space for users to share their thoughts and feelings, serving as a supportive partner for mental well-being.

Realtime voice is powered by Volcengine Doubao end-to-end speech model.
Docs: Doubao End-to-End Realtime Speech API.

For local development, run the backend with Go directly and use pnpm dev server for the frontend:

# Start backend (in project root)
go mod download
go run .

# Start frontend (in web directory)
cd web
pnpm install
pnpm dev

For local deployment using Docker Compose:

# 1. Build frontend assets
cd web
pnpm install
pnpm build

# 2. Start services with Docker Compose
cd ..
docker compose -f compose.dev.yaml up -d

The services will build Docker images locally and start containers. The web interface will be available at http://localhost.

Build Docker images and push to Docker Hub:

# 1. Build frontend assets
cd web
pnpm install
pnpm build

# 2. Build Docker images
cd ..
docker compose -f compose.dev.yaml build

# 3. Tag images
docker tag meow-ai-api:latest meow2149/meow-ai:api-latest
docker tag meow-ai-web:latest meow2149/meow-ai:web-latest

# 4. Push to Docker Hub
docker push meow2149/meow-ai:api-latest
docker push meow2149/meow-ai:web-latest

Deploy on production server:

# 1. Configure domain on Cloudflare and enable proxy

# 2. Prepare configuration files
# Copy config.yaml and compose.yaml to the server

# 3. Start services
docker compose up -d

Services will automatically pull the latest public images from Docker Hub and start. Watchtower will automatically monitor and update containers.

Meow-AI is MIT licensed.