Files are copied into a project workspace. The source files stay untouched.
STEM INTELLIGENCE / REVIEW-FIRST WORKFLOW
Know what every stem is before the mix begins.
Selah Tagger combines filename evidence, acoustic analysis and optional model signals to organize session files while keeping uncertain decisions visible for human review.
Technical beta. No calibrated probability claims. Human review stays in the loop.

Internal names are sanitized, verified and tracked through the project manifest.
The system surfaces ambiguity instead of hiding it behind a confident-looking number.
FROM FOLDER TO SESSION
Four steps. One reviewable result.
- 01
Upload
Bring WAV, AIFF, FLAC, MP3, OGG or M4A stems into a project workspace.
- 02
Analyze
Filename evidence, acoustic features, adaptive memory and optional model checkers contribute separate signals.
- 03
Review
Inspect alternatives and reasoning for files that need confirmation. Correct labels in canonical English.
- 04
Export
Download an organized ZIP, manifest and intelligence report without renaming the originals in place.
MULTILINGUAL EVIDENCE / CANONICAL OUTPUT
Understands the session language. Exports one clear taxonomy.
Filename matching can recognize English, Brazilian and European Portuguese, Spanish and French terms. Final category labels remain canonical English for predictable exports and integrations.
TECHNICAL BETA / NO CARD
Bring the messy folder. Keep the final decision.
Create your free account and experience the complete upload, analysis, review and export loop with one project.