tone-deaf
Hand it lyrics and a score, and it sings them: an AI voice on the melody, a synthesized band underneath.
The score is a single plain-text file: melody, lyrics under the notes, and one chord per bar. From those chords it arranges its own backing band (pop, ballad, or acoustic, with drum fills, a walking bass, and piano or guitar), while DiffSinger — an open model trained on real singing — sings every syllable on its note.
Before finishing, it checks its own work by transcribing the recording back and measuring every note's pitch. The name is a joke: pitch is the one thing it always gets right.

Listen
The work behind it
- Lines written
- 6,602
- Files
- 35
- Days worked
- 8
- Commits
- 35
- Started
- 9 Oct 2026
- Last change
- 19 Nov 2026
What it would have cost
$25,000to have this built for you (its answers ranged $25,000–$26,000)
$3,000what it is worth to own today
Niche pipeline, depends on third-party voicebank with non-commercial license; no users or revenue evident.
Priced on 15 Nov 2026 by an AI model that read the project's source code: the middle of 3 answers. How this is worked out.
All projects in the showroom · Ask for a seat
Measured 22 Nov 2026.
Work