Alright, I've hit my limit. I'm using Otter.ai to transcribe post-mortems and engineering syncs, which are heavy on action items and lists. The output is a formatting disaster. It consistently mangles numbered steps and turns bullet points into garbled run-on sentences.
I need this to feed cleanly into our docs and Jira tickets without manual rework. The "smart formatting" toggle is on, but it's clearly not parsing the speaker's intent.
Example of the problem:
* **What was said:** "Okay, next steps: one, roll back the deployment. Two, check the database connection pool. Three, add more logging to the API."
* **What Otter gives me:** "Okay next steps one roll back the deployment two check the database connection pool three add more logging to the API."
Has anyone cracked the code on this? I'm looking for:
* A specific phrasing pattern that triggers better list recognition (e.g., do you have to say "bullet point" first?).
* Any hidden settings beyond the basic UI toggles.
* If the answer is "post-process with a script," fine—share your regex or parser. I'd rather have a reproducible pipeline step than rely on Otter's black box.
This isn't a nice-to-have. For GitOps, clear, automated documentation of decisions is critical. Right now, I'm spending more time fixing formatting than analyzing the content.
- pp
pipelines > meetings