Skip to main content

Build and run the EchoKit server from source

In this guide, we'll walk through building and running the EchoKit server locally from the source code.

Prerequisites​

Before starting, make sure you have:

  • Rust installed
  • Model service URLs for Whisper, LLM, and TTS (with optional API keys)

Steps to Run the EchoKit Server​

1. Get the Source Code​

git clone https://github.com/second-state/echokit_server.git

2. Configure the Server​

Edit the config.toml file to set up your LLM services. Since EchoKit supports OpenAI-compatible AI services, you are free to choose the provider you prefer.

3. Build the Server​

cargo build --release

4. Run the Server​

# Enable debug logging
export RUST_LOG=debug

# Run the EchoKit server in the background
nohup target/release/echokit_server &

If everything goes well, you’ll see output like this in your terminal：

[1] 67869
appending output to nohup.out

Logs are saved in nohup.out.

Server URL​

The EchoKit server runs as a WebSocket server. Its URL will look like this:

ws://YOUR_WIFI_ADDRESS:8080/ws

Use this URL when connecting your EchoKit device to the server.

Do not use localhost:8080 or 127.0.0.1, because your EchoKit device needs to connect over the network — it can't reach localhost on your computer.