反谱

反谱 is a browser-based AI music and audio utility AImed at Chinese-speaking musicians, learners, teachers, and short-vIDEo creators. Rather than ACTing as a single-purpose generator, it bundles several everyday music tasks into one online workspace, so a user can move from a rough recording to a usable score, an isolated backing track, or a brand-new song without installing desktop software. All processing hAPPens in the cloud, which keeps the interface simple and lets the service update without user-side upgrades.

What it does

The centerpiece is automatic music transcription. You upload an audio or video file, and the system analyzes pitch, rhythm, key, and chords to produce Standard staff notation or numbered notation. For guitar and bass sources it can add tablature. Around that core, 反谱 adds vocal and instrumental stem separation, text-to-song generation, score-to-audio conveRSIon, and Voice Cloning for spoken or sung text. The platform accepts common formats such as MP3, WAV, M4A, FLAC, MP4, and MOV, and returns files in widely used formats that can be imported into arrangement or practice software.

Key capabilities

• AI transcription: Converts audio or video into staff and numbered notation, with tablature for plucked instruments such as guitar and bass.
• Stem separation: Splits a mix into vocals and accompaniment, or further isolates piano, drums, bass, and guitar tracks for targeted practice, covers, or video scoring.
• Song generation: Takes keywords, lyrics, and a chosen style to produce an original song track.
• Score conversion: Turns static notation files into playable, editable audio for rehearsal or reARRangement.
• voice cloning: Clones a timbre for reading text, lyrics, or teaching material, with built-in voice presets.

Who it is for

The tool is designed for people who work with music but may not have formal theory training: hobbyists learning a song by ear, instrumental students practicing against isolated parts, teachers preparing lesson scores, independent artists sketching arrangements, and short-video creators who need background music or covers. The Chinese-language interface and three-step upload-process-download flow lower the entry barrier, though very noisy or heavily mixed source files will reduce transcription and separation quality.

How to get started

1. Open the service in a browser; no CLIent download is required.
2. Choose a module such as transcription, vocal separation, multi-instrument separation, song generation, or voice cloning.
3. Upload a clean, playable source file in a supported format.
4. Wait for cloud processing, then download the notation, stem files, or generated audio.

Pricing and access caveat

Access appears to be free to try in the browser, with credit-based limits likely on heavier tasks such as long files or repeated separations; pricing and free quotas may vary and should be confirmed on the official site before relying on them. The service is Chinese-language and web-only, so mobile and desktop use both depend on a browser connection.