Make the archive
answer back.
A source-aware workspace for tracing lyrics, audio, and the conversations that shaped their meaning.
Archive index
Imported material stays bounded.
No external sessions are queried. Files you choose are indexed locally for this browser.
Inspect the recording.
Cross-song comparisons
Every claim keeps its label.
Jadox Archives separates what the author has actually said from what an AI or researcher inferred. Import the source, attach the claim, and keep uncertainty visible.
Suggest a research workflow ↗Help & FAQ
Jadox Archives is a private, evidence-backed workspace. It keeps song records, conversation context, audio observations, and uncertainty together without pretending that missing source material is known.
Return to the archive ↗What exists in the archive right now?
The seed archive includes context from the actual sessions for dsFnkshnL, Bodycam, Contact High, and Strive 4 Perfection, plus the locally available Swipe to the Next One source. Claims are labeled as author-confirmed, interpreted, or open.
How do I import a conversation?
Choose Import evidence, then select a ChatGPT export or conversation-shaped JSON. The app indexes the file locally in this browser. It does not reach into private ChatGPT history by itself.
How do I add audio?
In Audio Lab, choose the archive record, then drop a WAV, MP3, M4A, OGG, FLAC, or AAC file into Upload a recording. The recording is stored in this browser’s local IndexedDB and can be loaded again without re-uploading it on that device.
What does audio analysis measure?
The lab can inspect duration, sample rate, channels, peak and true-peak estimates, RMS loudness, loudness range, silence intervals, stereo correlation, spectral centroid, tempo, and section changes when the browser path supports them. These are signal diagnostics—not judgments about whether a beat sounds good.
Why can analysis differ between browsers?
Browser audio decoders and memory limits differ. The app uses a lighter playback-first path for Brave and a fuller analysis path where supported, including Vivaldi. If playback sounds wrong, compare the original file in a desktop player and try the browser’s native format path.
How do I review or export an audio report?
Load an attached recording, use the player or waveform to inspect it, and add timestamp marks for moments worth revisiting. Export report saves the measured diagnostics and marks as JSON; you do not need to play the entire recording first.
What does the provenance label mean?
Author confirmed means the source directly supports the claim. AI interpretation means it is a proposed reading. Open question means the archive needs another source, an author confirmation, or a closer audio/lyric inspection.
Where is my data stored?
The current MVP stores archive records in localStorage and audio attachments in IndexedDB for this browser/device. The private hosted site is protected by Cloudflare Access, but this is not yet a multi-user database with server-side per-user storage.
What should I do when a lyric or audio payload is missing?
Import the original conversation, lyric file, or recording and attach it to the matching record. Keep the open question visible until the primary source is available; do not fill the gap with a demo or an invented transcription.