What BananaScribe is
A transcription tool built for production audio and video. It turns multi-track set recordings — typically one lavalier microphone per person — into a single, time-stamped, speaker-labeled script, and does the actual transcription work locally, without requiring an internet connection.
- How It Works — architecture, the processing pipeline, and how multi-track/MXF audio is handled end to end
- Security & Data Handling — what's local vs. networked, access control, storage/retention, updates, and disclosed tradeoffs
- Installation — system requirements, the installer walkthrough, updating, uninstalling
- User Guide — importing files and MXF takes, the dashboard, search, the viewer/player, exporting
- Administration — running BananaScribe for a team on a local network, accounts, sessions, operational notes
- FAQ & Glossary — terms (MXF, diarization, bleed suppression, turns...) and common questions
01Who it's for
Production and post teams working with multi-channel set audio — for example, MXF recordings carrying one isolated microphone track per cast or crew member — who need a readable, searchable transcript of what was said, by whom, and when, without sending footage to a third party to get it.
02What goes in, what comes out
Raw tracks
One audio channel per microphone/person, as recorded on set — MXF takes, or any individual audio/video file.
One script
Chronological, speaker-labeled, time-stamped text — exportable as plain text, Word (.docx), or PDF, per take or merged across several.
03Two ways to run it
| Desktop app | Self-contained Windows application on a single machine. Nothing to configure; one person, one install. |
|---|---|
| Shared instance | The same software, run once on a production's own local server, with a small team signing in from their own machines over that local network. See Administration. |
In both cases, the software and the content it processes stay on hardware the production controls.
Looking for the security answer specifically?
Go to Security & Data Handling — it covers offline processing, the one network dependency, access control, storage, update integrity, and tradeoffs we chose deliberately and are disclosing rather than hiding.