How 102 Legal Audio Files Built Our Bulk Service
We got an email in early February from a paralegal at a mid-size law firm. She had 102 audio files she needed transcribed — field recordings, phone calls, and meeting recordings related to a highway construction dispute case. Depositions with contractors. Recorded calls between project managers. On-site audio from inspections.
She had already tried our standard upload page. One file at a time, pay per file, download, repeat. After three files she wrote to us: "Is there any way to upload these all at once? I have 102 of them and doing this one by one is not realistic."
That email changed our product roadmap.
Quick Navigation
- The Request That Started Everything
- Why One-at-a-Time Doesn't Scale
- Building the Bulk System
- The Silent File Problem
- Catching Duplicate Filenames
- Volume Pricing: How It Works
- What the Self-Service System Looks Like Today
The Request That Started Everything
The paralegal — we'll call her Rachel — was working under a deadline. Her firm was preparing for mediation in a dispute over a state highway widening project. The general contractor alleged the subcontractor had deviated from grading specifications. The subcontractor claimed the specs were changed verbally during site meetings that were never documented in writing.
Both sides had recordings. Lots of them. Site inspection walk-throughs recorded on phones. Conference calls between engineers. Depositions taken over video. Planning meetings with municipal officials. Rachel's attorneys needed searchable transcripts of all of it to build their timeline of who said what, and when.
102 files. Ranging from 4 minutes to just under 2 hours each.
At the time, BrassTranscripts only offered single-file uploads. Upload one MP3 or MP4, get your transcript, pay, download, next file. That works fine for a podcaster processing one episode, or a journalist transcribing a single interview. It does not work for someone staring at 102 files with a mediation date approaching.
Why One-at-a-Time Doesn't Scale
We did the math on what Rachel's workflow would look like with our single-file system:
- 102 separate uploads — each requiring file selection, upload wait, and processing
- 102 separate payments — each at the standard $6.00 rate for files over 15 minutes
- 102 separate downloads — each transcript retrieved individually
- Total cost at standard pricing: potentially $612.00 with no volume consideration
Nobody should have to click "Upload" 102 times. The cost adds up without any volume discount. And the workflow of managing 102 individual transactions is its own full-day project.
Rachel needed a dashboard. She needed batch upload. She needed to process everything at once, pay once, and download a ZIP file.
So we built it.
Building the Bulk System
We initially built Rachel's batch as a managed service — we handled the upload and processing on our end. But the request pattern kept repeating. Research teams with 40 interview recordings. Call centers with a week's worth of phone calls. Podcast networks with a back catalog.
Every time, it was the same need: "I have a lot of files. I can't do them one at a time."
So we turned the managed service into a self-service product. Today, anyone can sign up at brasstranscripts.com/bulk/signup, verify their email, and get a private dashboard link. Upload files, start processing, pay when everything is complete. No approval process, no waiting for us to set anything up.
The system processes files concurrently — not one after another, but multiple files transcribing simultaneously. The same AI transcription engine and automatic speaker identification that powers every single-file job. Same quality, just more throughput.
The Silent File Problem
Here's something we didn't anticipate until Rachel's batch: not every audio file contains speech.
Two of her 102 files were recordings from a highway-mounted camera system. The camera had audio enabled, but the microphone only captured road noise — traffic, wind, and the hum of construction equipment in the distance. No human voices. No words.
Our transcription engine processed them and returned zero words. In a one-at-a-time workflow, that's just a failed file you shrug off. But in a batch of 102, charging someone for files that contain no speech felt wrong.
So we built silent file detection into the system. When a file completes processing with zero words detected, it gets flagged automatically:
- Status shows "No speech" on the dashboard
- Not included in the invoice — you don't pay for it
- Doesn't count against your file limit — silent files are excluded from your account's file count
Rachel's two highway camera recordings were flagged immediately. Her invoice reflected 100 billable files, not 102. She didn't have to dispute the charge or contact us. The system handled it.
This matters more than you might think. Research teams sometimes include ambient room recordings or equipment tests in their upload batches. Podcasters occasionally include intro music tracks or sound checks. Any batch of real-world files is likely to include a few that have no transcribable content, and nobody should pay for those.
Catching Duplicate Filenames
Rachel's files arrived with names like Inspection_Recording_Site_4.mp3 and Deposition_Martinez_2025.mp4. But she also had two files both named Meeting_Notes_January.mp3 — one from a planning meeting and one from a site meeting, saved with the same filename on different dates.
Uploading two files with identical names is an easy mistake. Different folders on your computer, same filename. You don't notice when you're selecting 50+ files at once.
Our dashboard catches this before processing starts. When you add files, the system checks for duplicate filenames (case-insensitive) within your current batch. Duplicates get flagged with an amber "Duplicate" badge and a summary notice at the top of the dashboard telling you how many duplicates were detected.
You can review each flagged file and remove the actual duplicate before starting processing. This prevents paying twice for the same recording — or worse, getting two identical transcripts and not realizing it until you've already paid.
For Rachel, the duplicate flag caught the two Meeting_Notes_January.mp3 files immediately. She renamed one, re-uploaded it, and processing started clean.
Volume Pricing: How It Works
Standard BrassTranscripts pricing is $2.50 for files up to 15 minutes and $6.00 for files 16–120 minutes. That's designed for occasional use — a few files per week.
Bulk pricing works differently. It's per-file, with automatic volume tiers:
| Batch Size | Price Per File |
|---|---|
| 1–19 files | $90.00 flat (20-file minimum) |
| 20–49 files | $4.50 per file |
| 50–99 files | $4.00 per file |
| 100–249 files | $3.50 per file |
| 250+ files | $3.00 per file |
Rachel's 100 billable files (after the two silent files were excluded) hit the 100–249 tier at $3.50 each. Total: $350.00.
Compare that to processing 100 files individually at $6.00 each: $600.00. The bulk pricing saved her firm $250 on a single batch.
Every file includes speaker identification at no extra cost. Failed and no-speech files are not billed. The tier applies automatically based on your batch size at checkout — no coupon codes, no negotiation, no sales call.
What the Self-Service System Looks Like Today
What started as one paralegal's emailed request is now a fully self-service product. Here's how it works:
1. Sign up — Create an account at brasstranscripts.com/bulk/signup. Verify your email.
2. Get your dashboard — You receive a private link to your bulk dashboard. Bookmark it. Access links are valid for 10 days and you can request a fresh one anytime at brasstranscripts.com/bulk/login.
3. Upload files — Drag and drop or select files. All 11 supported audio and video formats accepted (MP3, MP4, M4A, AAC, WAV, FLAC, OGG, Opus, WebM, MPEG, MPGA). Maximum 250MB per file, up to 2 hours per recording.
4. Review before processing — The dashboard flags duplicates and shows your file count. New accounts start with a 100-file limit, which increases automatically after each payment (200 after first payment, then +100 per subsequent payment).
5. Start processing — Click the button. Files process concurrently with automatic speaker identification and language detection across 99+ languages.
6. Pay when complete — Once all files finish, checkout through our secure payment system. Silent files and failed files are excluded from your invoice automatically.
7. Download — Get all your transcripts as TXT files. Transcripts are available for download for 10 days after payment.
The whole thing runs without us in the loop. No emails back and forth, no custom quotes for standard batches, no waiting for business hours.
Rachel's firm won their mediation. We don't know how much the transcripts contributed — that's attorney work product, not ours. But we do know that 100 searchable transcripts with speaker labels gave her attorneys a way to build a timeline that would have taken weeks to construct by listening to recordings manually.
Her request taught us something we should have realized sooner: when someone has a lot of files, the biggest barrier isn't price or quality. It's workflow. Make the workflow match the scale of the problem, and everything else follows.
If you're sitting on a batch of recordings that need transcribing — whether it's 20 files or 200 — the bulk transcription guide walks through the complete process. Or just sign up and start uploading.
Related Resources:
- Bulk Audio Transcription: Transcribe 20+ Files - Complete bulk service overview
- Speaker Identification: Complete Guide - How automatic speaker labeling works
- Interview Transcription for Qualitative Research - Research-focused transcription workflows