search

chochinlu / suno-song-generator

6 stars 2 forks source link

readme

Suno Song Generator

This Suno Song Generator is designed for users who are not necessarily music professionals (like myself).

Have you ever listened to a song and thought, "Wow, this song is amazing! I wish I could create something like this," but you lack the knowledge of music production, genres, or instrumental details?

This tool allows you to:

  1. Provide a YouTube link of a song you like or input related lyrics
  2. Utilize AI analysis to generate a similar style and lyrics to the song you love
  3. Customize the output to your preferences
  4. Finally, use the Suno API to produce a song

With this generator, you can turn your musical inspiration into reality, even without extensive musical expertise.

How to run

Register a Suno account

First, you need a Suno account. Register at Suno.

Langfuse

Langfuse is an open-source LLM engineering platform that helps teams collaboratively debug, analyze, and iterate on their LLM applications.

The app has added the langfuse observer decorator to the functions that interact with the OpenAI API.

You can use the self-hosted LangFuse to observe the tracing data. Flow this repository and run docker compose up -d to start the LangFuse server.

Then, you can view the tracing data at http://localhost:3000/

Start the Suno API server

You need to start the Suno API server, please refer to the Suno API documentation

Note: You need to create a .env file at the root of the Suno API server and add the following environment variables:

SUNO_COOKIE=<your_suno_cookie>

How to obtain the cookie of your Suno account

Note: If you are using the Langfuse server mentioned below, as it occupies port 3000, you need to specify a different port for the Suno API server, for example: http://localhost:4000

I recommend using docker to run the Suno API server. Before running the docker, you need to change the docker-compose.yml file to specify a different port from the one used by the Langfuse server.

For example, change the docker-compose.yml file to:

    ports:
      - "4000:3000"

Then, run the docker compose: 

docker compose build && docker compose up -d

Run the Song Generator

Create .env file and add the following environment variables:

OPENAI_API_KEY=<your_openai_api_key>
SUNO_API_HOST=<your_suno_api_host>

Fllow the instruction in the langfuse server, sign up for an account, create a new project, and get the LANGFUSE_SECRET_KEY and LANGFUSE_PUBLIC_KEY, add them to the .env file: 

LANGFUSE_SECRET_KEY=<your_langfuse_secret_key>
LANGFUSE_PUBLIC_KEY=<your_langfuse_public_key>
LANGFUSE_HOST="http://localhost:3000"

Create virtual environment (recommend) and install dependencies:

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Start the Song Generator:

python app.py

The web app will run on http://localhost:7860

Screenshot

image1

image2

image3