Skip to content

USER-GUIDE.md

Latest commit

 

History

History
94 lines (71 loc) · 3.9 KB

USER-GUIDE.md

File metadata and controls

94 lines (71 loc) · 3.9 KB
Raw

User Guide

App: https://app-metadata.hardbanrecordslab.online

1. Account & credits

  1. Open the app, choose Sign Up, register with email + password (or Google).
  2. New accounts get 3 free credits. 1 credit = 1 full analysis. Exporting the results of an analysis you already ran is free.
  3. Out of credits? Settings → Buy credits → pick a pack → Stripe Checkout. Credits appear as soon as the payment confirms (you're redirected back with ?billing=success).
  4. Have a promo/referral code? Settings → Redeem code.

2. Analyse a track

  1. From the dashboard, New Analysis (or drag a file onto the dropzone).
  2. Supported: MP3, WAV, FLAC, AIFF — up to ~100 MB.
  3. Choose the mode:
    • Flash — faster, 2-model AI consensus.
    • Pro — adds a third independent model for a genuine 3-way vote.
  4. Watch progress in real time; the written description streams in as it's generated.
  5. A run takes roughly 20–45 s depending on mode and track length.

3. Read the results

Results are grouped into cards:

Card What it shows
Track Identity Title, artist, detected language, era/production style
Sonic Analysis BPM, key/mode, energy, loudness (LUFS), brightness, dynamics, spectral profile
Classification & Style Main genre, sub-genres, moods, instrumentation, vocal style — each with a confidence meter
Structure Section/arrangement read from the signal
Identification Matches from ACRCloud / AcoustID / MusicBrainz (if the recording is known)
External Data Spotify / Last.fm / Discogs enrichment
Marketing Keywords, use-cases, target audience, similar artists, the prose description
Commercial / Legal Sync-placement suggestions and the AI-content caveat
Copyright Authenticity DNA (SHA-256), fingerprint
Visuals AI cover-art suggestion + colour palette
Validation Report Consistency checks (e.g. BPM vs genre) and confidence summary

Every AI tag has a confidence score. If you disagree, override it — the model's answer is a suggestion, not a verdict (see the AI Disclaimer in the app footer).

4. Export

After an analysis, export in any format — no extra credit:

  • MP3Tag CSV — bulk import into MP3Tag.
  • JSON — the full structure with ISO timestamps.
  • DDEX ERN 4.3 — XML for delivery to DSPs via a distributor.
  • CWR 2.1 — Common Works Registration for PRO/CMO submission.

You can also write tags straight back into the audio file (Tools → Tag), strip metadata, or convert format.

5. Certificate of Authenticity

Generate a signed certificate (PDF) for a track. It embeds the Authenticity DNA hash and is pinned to IPFS, with a public verification page at /verify/<id> that anyone can check. Useful as a pre-distribution provenance record.

6. Batch

Batch Processor analyses many files at once (a queue with a few parallel jobs). Each file still costs 1 credit. Good for auditing an album or a back catalogue.

7. History & settings

  • History — every past analysis, re-openable and re-exportable.
  • Settings — profile, credits & purchases, redeem codes, theme (light/dark).

8. Verifying someone else's certificate

Open /verify/<certificate_id>. The page shows the certified metadata, the Authenticity DNA, the IPFS link, and whether the hash still matches — no account needed.

Troubleshooting

Symptom Likely cause
"Zaloguj się, aby skorzystać" Not signed in.
"Brak kredytów" (HTTP 402) Out of credits — buy a pack or redeem a code.
Genre comes back generic / "Pop" fallback Too few AI models answered; the DSP-only fallback ran. Retry, or use Pro mode.
Identification card empty The recording isn't in the public databases — normal for unreleased tracks.
Analysis stuck Refresh; the job is in the queue and will still complete server-side.