Skip to content
Notifications
Clear all

How do I batch edit speaker names across 50 transcripts?

1 Posts
1 Users
0 Reactions
78 Views
(@ide_tinkerer)
Reputable Member
Joined: 7 months ago
Posts: 338
Topic starter   [#4215]

Okay, so I've hit a classic automation problem that I'm hoping someone here has already navigated. I've been using Otter.ai for a few months to transcribe team retrospectives and technical interviews, and I've amassed a library of about 50 transcripts. The speaker labeling is… inconsistent. Sometimes it's "Speaker 1," sometimes it's a misheard name like "Carrie" instead of "Kerry," and sometimes it's just initials I typed in a hurry.

Manually opening each transcript in the web interface and clicking through to rename speakers is a level of tedium I refuse to accept. I need to batch edit these names. Has anyone figured out a workflow for this?

My ideal scenario would be something like:
* Export all transcripts in a structured format (JSON, CSV, maybe SRT/VTT?).
* Run a script to find/replace speaker labels based on a mapping file (e.g., "Speaker 1" -> "Alex (Dev)", "Carrie" -> "Kerry (PM)").
* Re-import the corrected transcripts, or at least generate a corrected version I can store alongside the original.

I've poked around the Otter.ai web UI and the export options. Here's what I see so far:

**Exports Available:**
* TXT (plain text, loses speaker labels)
* PDF (formatted, but not machine-friendly)
* DOCX (Word doc, might retain some structure?)
* SRT / VTT (caption files, which *do* include speaker labels as part of the cue text)

The SRT/VTT export seems the most promising for automation. A snippet looks like this:

```vtt
WEBVTT

00:00:01.210 --> 00:00:04.730
Speaker 1: So the initial spike estimate was three days.

00:00:04.730 --> 00:00:07.920
Carrie: We hit a snag with the API rate limits though.
```

If I could get all 50 transcripts as VTT, I could write a Python script (or use `sed`/PowerShell) to do a global replace. But before I go down that path:

1. **Is there a hidden bulk action in the Otter.ai dashboard I've missed?** A "Manage Speakers" page that applies across conversations?
2. **Has anyone successfully scripted this using the Otter.ai API?** I glanced at the API docs—it seems focused on creating new transcripts, not batch-editing existing ones.
3. **If the VTT method is the way, what's the best practice for re-associating the cleaned file with the original transcript in Otter?** Or do you just keep the corrected VTTs as your source of truth outside the system?

My secondary concern is preserving the edit history within Otter for future corrections, but batch accuracy is the immediate priority. I'm comfortable with command-line tools and scripting, so don't shy away from technical solutions.


editor is my home


   
Quote