Show HN: Sonauto – A more controllable AI music creator

Original Article ↗ Hacker News Discussion ↗

AI-generated music tools like Sonauto, Suno, and Udio are rapidly improving, letting users create full songs from text prompts with minimal input, but many listeners still find the results low-fidelity, emotionally flat, or overly generic. Commenters praise Sonauto’s ease of use and emerging controls (rhythm guidance, planned variation and DAW‑style features), yet argue the real value will come from tools that integrate with professional workflows, output stems, and enable fine-grained iteration rather than one-shot “slot machine” tracks. Alongside technical comparisons, the thread raises concerns about copyright, artist livelihoods, login requirements, and whether infinite on-demand music will complement or undermine human-made art.

Overall Reception

  • Many find the tool fun, impressive for minimal input, and less painful to use than competitors.
  • Others are underwhelmed by quality, emotion, and “sameness” of vocals; some describe it as “washed out,” “bloodless,” or “elevator music.”
  • Several commenters note this is early days and “full songs from prompts” are the “hello world” of the tech.

UX, Features, and Control

  • Users appreciate hidden “advanced dials” and quick generation, but some find the interface confusing (e.g., only last input editable, required song title).
  • Key differentiator advertised: more control via “Rhythm Control,” BPM assist, and upcoming “variations” that re-use parts of an existing generation.
  • Some report bizarre outputs (no backing music, one‑note songs) possibly linked to BPM assist; suggestion to try “instrumental” options.
  • Strong demand for finer control: stems per instrument, partial edits (e.g., add verse, change drums), extend/iterate without “rerolling” the whole track.

Comparison to Other Tools (Suno, Udio, etc.)

  • Many compare directly to Suno and Udio:
    • Udio often praised for “songwriting/performance,” though described as safe, “elevator‑like.”
    • Suno seen as better in some cases but with robotic/metallic voices and lower average quality; cherry-picking required.
  • Some say Sonauto currently lags in prompt adherence, voice variety, and clarity; others note specific strong outputs (e.g., certain genres, languages).

Integration with DAWs and Pro Workflows

  • Multiple musicians and producers want:
    • Exportable stems, MIDI, and chord progressions.
    • VST/DAW integration and “AI as bandmate” rather than “infinity jukebox.”
    • Tools to clean up live recordings, add bass/drums, and iterate arrangements.
  • Several argue the real opportunity is assisting real producers, not just casual meme songs.

Authentication and Access

  • Heavy pushback on “Sign in with Google” as the only option; concerns include privacy, dependency on an unrelated platform, and professional users avoiding social logins.
  • Some argue most mainstream users prefer Google login; others insist basic email/password is easy and should be added.
  • A few note confusion between “free and unlimited” claims and the requirement to log in via Google.

Ethics, Copyright, and Impact on Musicians

  • Repeated questions about training data sources, licensing, and legal risk given litigious music industry and plagiarism rules.
  • Debate over whether AI music will:
    • Mainly be cheap background/marketing audio, or
    • Compete with human-created music on streaming.
  • Strong split views:
    • Some see AI as a tool that empowers musicians, speeds “80% familiarity” so humans can focus on the “20% novel.”
    • Others see it as depressing, “fake emotion,” job-threatening, or culturally harmful; some object to replacing human hobbies with automation.

Open Source and Future of the Space

  • Requests for open models; developers indicate this conflicts with funding/compute realities.
  • Many expect a future “Stable Diffusion for music”: open, VRAM‑friendly, with fine-tuning and control-net–like extensions.
  • Commenters anticipate the space rapidly filling (Suno, Udio, Sonauto, Stability, etc.); consensus that the long-term differentiator will be product and workflow integration, not just raw model capability.