CornyBot Voice Lab

Turn any lyrics into your voice — trained from 13 minutes of audio. Forensic match: —%.Train your model to unlock your forensic score.

Prepare Training Data

The 3-Stage Voice Pipeline TTS + SVC + Cadence

Every voice conversion passes through three specialized stages. Click each stage to learn more.

Stage 1

Edge-TTS

Microsoft Neural TTS

AI reads your lyrics with natural prosody and rhythm

Stage 2

RVC v2

HiFi-GAN + ContentVec

Your voice model transforms the speech into your unique timbre

Stage 3

Cadence Mapping

BPM-aware alignment

BPM-aware rhythm alignment matches your natural rap cadence

💡 Architecture: RMVPE Pitch Extraction

RVC v2 uses RMVPE for F0 extraction, outperforming crepe and harvest methods. Combined with ContentVec speaker embeddings, this enables high-fidelity voice conversion from just 13 minutes of training data.

f0_method=rmvpe | embedder=contentvec | sample_rate=40000

Training Data

—

Current Epoch

—%

Forensic Match

⚡

Your voice model is R&D-qualifying IP

Every training run, render, and quality measurement on this page is auto-classified as an IRC §41 qualified research expense. When forensic match crosses 80%, the model becomes mintable on Solana as a Token-2022 IP asset with revenue-share.

View Readiness Score How R&D Mint Works →

Quality Metrics Loading...

—%

Composite Match

—%

Spectral Envelope

—%

Pitch Similarity

—%

Energy Profile

—%

Warmth Match

—%

Harmonic Structure

Voice Model Gallery 1 Model

Training History —

Voice model quality trajectory across training iterations.

Guided Demo No Auth Required

Experience the voice conversion pipeline. No GPU needed — runs on our cloud infrastructure.

See Sample Lyrics

View lyrics from the UNICORNY catalog

Convert

Pipeline processes TTS → RVC v2

Compare

Hear before/after waveforms

See Metrics

Quality scores + waveform analysis

Demo Sample Lyrics

We been gone but now we're back in it
Stack it up, no cap, we're racking it
Every line we write is automatic
Pull up on the scene, it's cinematic

Community Voice Contributors

Artists who contribute vocal stems earn $CORNY tokens when the model is used.

Contributors

Clips

Minutes

$CORNY Distributed

Loading contributors...

Contribute Your Voice

Upload vocal stems to help train the model. Sign consent, get credited, earn $CORNY tokens when your contributions improve the voice.

Upload Stems

Try with Your Own Lyrics

Generate a verse and hear it in your voice. $0.005/second — 73x cheaper than ElevenLabs.

Open CornyBot Studio

Or integrate via API →