MuseTree is a custom front-end for OpenAi's MuseNet, the AI music generator. The official app is a toy designed to show off the research. In contrast, MuseTree is designed for real music production, and has been built from the ground-up with that in mind.
MuseTree is still in active development. I'm very happy to take suggestions, so please open a GitHub issue or comment on an existing one with your ideas! If you want to chat with me in real time, I live-code on twitch.tv Mon-Fri 10:30am-3pm BST. At the minute, those streams are mostly my work on MuseTree, so come along and say hi!
The latest version of MuseTree is always available online at http://musetree.stevenwaterman.uk
- Full tree support:
- Generate more than 4 options
- Develop multiple tracks at once
- More generation options:
- 604 genres available
- Generate up to 1000 tokens at a time
- Set the max response length to prevent MuseNet errors and generate compositions of unlimited length
- Change the temperature to escape repeating patterns or overfitting to real songs
- Full persistence options:
- Save/Load the full tree to work on later
- Export
wavaudio - Import/Export
midi
- Fixes some MuseNet bugs
- Notes that last forever
- Failed requests automatically retry
- No length limit
MuseTree uses the svelte web framework with TypeScript.
Quick Start:
git clonenpm inpm run dev
Contributing:
MuseTree is still in an early state. Any and all contributions are welcome! Please have a look at the issue list and have a go at fixing one, or add more! I'm friendly, I promise.]
Using Docker
Using it with Docker is as easy as running this command:
docker run -p 8080:80 skillcrafter/musetree
It is based on the httpd:2.4 image. Future Plans are to support Nginx and some other Webserver.