# Run locally

The preferred way to run the project is to download the source code and run it locally. This way, you can customize the bot to your liking and you don't have to worry about the servers going offline.

You need to have node v16+ and git installed on your machine. You can download them from the following links:

• Node.js
• Git

The project should work on Windows, Linux and MacOS. If you have any issues please report them in the github issues page.

First, you need to clone the source code. You can do this by running the following command:

git clone https://github.com/miku-gg/miku.git

Next, you need to run make install to install the dependencies. You can do this by running the following commands:

cd miku
make install

This command will check the dependencies and install the required packages for npm and pip.

Next, you will be prompted for the environment variables

• AZURE_API_KEY, a valid API for querying Azure TTS models
• ELEVENLABS_API_KEY, a valid API for querying ElevenLabs TTS models

Backend API keys and URLs are set in the bot directory config. You can take a look at the How to use the endpoints guide for getting one of these values.

After setting up the .env file, you can run the project by running the following command:

make run

This should run up 6 servers:

http://localhost:5173 - the web app for interacting with the bots (NodeJS)
http://localhost:8484 - the services server for querying the external services (NodeJS)
http://localhost:8585 - the bot directory server for installing and listing bots (NodeJS)
http://localhost:8586 - the bot builder server for creating and editing bots (NodeJS)

Your initial place should be http://localhost:8585 to install the bots. It should open automatically in your default browser.