How to transcribe a meeting recording without buying Otter
Whisper open-weights are good enough now that you don't need a $17/mo subscription for the 80% case. Here's the workflow plus when Otter still wins.
Otter.ai is good at what it does — live meeting capture with calendar integration. If you live in back-to-back meetings, the $17/month is well spent. But the 80% case for most people is async transcription: you have a recording, you want a transcript. That doesn't need a real-time meeting assistant — it needs Whisper.
Why this works now
OpenAI's Whisper large-v3 model (released 2024, refined through 2025) hits word-error rates under 5% on clear English speech — better than typical human transcribers. Combined with Pyannote for speaker labelling, you get a fully labelled transcript for any recording you can upload.
The 60-second workflow
- Export the recording from Zoom / Teams / Meet (audio-only preferred).
- Open the Apexkit transcribe tool.
- Upload — drag-and-drop or click to pick.
- Pick language (auto-detected if mixed) and toggle speaker labels for Pro+.
- Click Transcribe. A 60-min meeting takes 3-4 min on the standard queue, ~90s on Pro+ priority.
- Download as SRT (with timestamps) or TXT (plain).
When Otter still wins
- Live meeting note-taking— Otter auto-joins your meetings and surfaces a transcript+summary email after. We don't do live yet.
- Searchable historical archive— Otter's decade-old library + tagging is much more polished than ours.
- Action-item extraction— Otter pulls follow-ups out of meetings automatically. We have a manual “summarise + extract action items” workflow but it's 2 clicks, not 0.
Cost math
Otter Pro is $17/mo with a 1,200-minute cap. The same minutes via our Pro tier cost roughly $0.40 in API spend — and you don't pay extra for it because it's bundled into your £17.99/mo Apexkit Pro. If audio is > 30% of your AI usage, the math works out; if it's < 10%, either tool fits.
Honest full comparison on the dedicated Apexkit vs Otter page.