Dadabots logo

Dadabots

Eliminate humans from music

4.6
Try Dadabots

Open Source Voice Agent SDK

Integrate voice into your apps with VideoSDK's AI Agents. Connect your chosen LLMs & TTS. Build once, deploy across all platforms.

Upvote Now

Overview

Hero
Get Started
Dadabots is a hybrid of band, hackathon team, and research lab where musicians fascinated by mathematics develop neural audio networks to generate music. Their systems create continuous, ever-evolving music livestreams and albums using deep learning, blurring the lines between scientific research, artistic collaboration, and technological innovation across genres like death metal, free jazz, and more. By writing code, publishing research, and collaborating with artists, Dadabots is at the forefront of AI-generated music.

How It Works

  • Deep Neural Networks: Uses architectures like SampleRNN to generate raw audio music.
  • Training Data: Neural nets are trained on acoustic waveforms from existing music or albums.
  • Prediction Process: The AI repeatedly predicts the next audio fraction of a millisecond, learning through millions of cycles.
  • Music Generation: After training, the AI “hallucinates” new music hours at a time.
  • Curation: Custom tools help select and arrange generated audio into albums.
  • Hyperparameter Tuning: Intensive trial and error with network size, learning rate, hierarchy, and optimizers.
  • Unsupervised Learning: The model receives no rules about music theory or MIDI, working exclusively with raw sound.

Use Cases

Continuous Music Livestreaming
Power a 24/7 livestream of AI-generated music in select genres like death metal or DnB, offering unique, evolving listening experiences.
AI-Powered DJing and Performance
Utilize Dadabots' technology for live performances and DJ sets, generating real-time music or infinite remixes on stage.
Artist Collaboration & Inspiration
Enable human artists to collaborate with AI and use it as a creative tool for new sounds, genres, and experimental projects.

Features & Benefits

  • Raw Audio Neural Networks: Captures detailed acoustic timbres beyond MIDI
  • 24/7 Livestream Music Generation: Infinite, evolving AI music channels
  • Multi-Genre AI Music Creation: Covers diverse genres like death metal, mathcore, breakcore, jazz
  • Music Deepfakes & Imitation: Generates music inspired by artists (e.g., Nirvana, John Coltrane)
  • Scientific Research & Publication: Publishes findings and open-source tools (SampleRNN, dadaGP)
  • Artist Collaboration Opportunities: Works with bands/artists (e.g., Reeps ONE, Lightning Bolt)
  • Human Augmentation Focus: Develops tools for users to sculpt AI music by taste
  • Open Comprehension & Accessibility: Promotes understanding and democratization of AI music tools

Target Audience

  • Musicians and artists keen on experimental, avant-garde, or extreme genres
  • Researchers and scientists in AI, deep learning, and audio synthesis
  • Music listeners seeking new, continuously generated audio experiences
  • Indie creators and youths interested in AI-powered music production
  • Those fascinated by the fusion of math, technology, and art

Pricing

  • Dadabots is available for free to users.
  • No cost required to access its music streams or technology.

FAQs

What is DADABOTS?

Dadabots is a hybrid of a band, hackathon team, and research lab. The group consists of musicians passionate about mathematics, involved in scientific research, engineering software, making music, and collaborating with other artists.

How did you get started working on DADABOTS?

Founders CJ Carr and Zack Zukowski met in 2012 at Music Hack Day MIT. They began by making remix bots that posted hundreds of songs per hour to Soundcloud until banned, inspired by the idea of machines generating art.

What inspired you to program an AI to create/replicate music?

They were inspired by advancements in deep neural networks, especially image style transfer and Wavenets’ voice synthesis, motivating them to model musical style and generate sound with target timbres through AI.

How does creating music with SampleRNN work?

Dadabots uses the original SampleRNN research code, training hierarchical LSTM networks on waveforms from albums. The AI predicts micro-slices of music, with humans curating and arranging the output into albums. Extensive hyperparameter tweaking is involved.

What’s the difference between Dadabots’ approach and other computer-generated music (like MIDI)?

Dadabots uses unsupervised training on raw audio, without rules or knowledge of music theory. This direct approach captures the full richness and timbre of real music, unlike MIDI, which only encodes discrete notes and lacks nuances.

Do you plan on using this software for retail purposes?

Dadabots anticipates that neural synthesis will feature in future audio software. As hardware improves and algorithms get more efficient, users may eventually instruct their DAWs to generate new music styles on demand.

Featured Products