The Rap Mixer

What the artist makes × what the context rewards → several outcomes, not one universal score

Scores explain a configuration in a selected context. They are not objective artistic truth.

Speech-to-text
Semantic analysis
Credential source

This public app has no shared or owner API keys. Use the free local modes, or paste your own provider key below — it is held automatically in server-side session memory and used by all tabs. Cloud calls also require the consent box.

Cloud sharing notice: depending on the provider, audio, transcript text, or extracted features may leave this server. Local modes send nothing.

Enter providers independently. Both keys can be held at the same time; they never enter serialized Gradio state.

Lyrics source
Transcribe first, review the visible text, then select it for scoring.
Time signature
Blend contexts
Cypher
0 100

Bank A — artist-controlled performance

0 100
0 100
0 100
0 100
0 100
0 100
A input mode

Bar analysis table

Inverse analysis — improve outcomes while protecting identity

Targets (all selected by default)
Rap membership
Intelligibility
Musicality
Potency
Replay depth
Cultural resonance
Battle effectiveness
Emotional impact
Innovation
Commercial accessibility
Artistic distinctiveness
Audience fit
Protected A dimensions
Protected outputs
1 20
1 25

Prioritized recommendations

Prototype coefficients and cultural interpretations are configurable assumptions, not empirically validated truth. No audio or transcripts are retained unless you explicitly export.