no bot in the call
the entire meeting-bot category — otter, fireflies, fathom, read.ai — joins your zoom or google meet as a participant to record and transcribe in real time. that's a different product, with different tradeoffs:
- it announces itself in the call. "[bot name] has joined." half your interview subjects don't want that. some organizations forbid it.
- it records on the bot's network, not yours. the bot vendor holds the audio. for sensitive conversations that's the wrong posture.
- it works on the meeting platforms it has integrations for. riverside, squadcast, logic, audacity, zencastr, a field recorder — none of those have a "let the bot in" toggle.
we don't do bots. you record the meeting on the platform you already use, in whatever quality you already capture, and you upload the file when it's done. for podcasters and writers who record conversations to turn into copy, this is the right shape.
the workflow
- record locally. zoom (cloud or local recording), riverside or squadcast (per-track audio), logic or quicktime (your own field recorder feed), google meet (with the host's recording on), or a separate handheld next to your laptop. whatever you're already doing for your podcast or interview recording, keep doing.
- upload the file. mp4, mov, m4a, mp3, wav. up to 5 GB per file. for video meetings, drop the .mp4 in directly — we extract audio on our end. for multitrack recordings (riverside, squadcast) either upload each speaker track separately for cleaner speaker separation, or upload the mixdown.
- transcription runs. on a 60-minute meeting, the first pass is ready in 1–3 minutes (cloud mode) or roughly real-time (on-device private mode for confidential conversations).
- fix speaker labels. "speaker 1" becomes the host. "speaker 2" becomes the guest. propagated through every turn at once. for multitrack uploads the labels are usually right on the first pass.
- extract what you actually wanted. highlight action items, decisions, quotable lines, interesting tangents. each highlight exports with a timestamp link back to the audio.
- export by job. an action-items list (markdown checkboxes) for your task manager, a clean transcript for the project archive, a highlighted-quotes file for the blog post or podcast episode, .docx if it has to go to a client.
three shapes of meeting, one transcript
for podcasters, bloggers, and authors recording meetings to turn into copy, the same recording usually has to do three jobs:
- action items. the things you agreed to do or to follow up on. usually 6–10 lines on an hour call. extracted into a checklist you paste into your task manager.
- quotable lines. the sentence the guest said that's going on the promotional graphic, the tweet, the lede of the post. verified by clicking the word — hear the guest say it — before publication.
- blog draft starter. the meeting itself, lightly edited, becomes the first draft of the post. fillers removed, paragraphed at thought boundaries, attributions clean. a one-hour conversation yields a 1,500–2,500 word draft you cut down rather than write up.
verify quotes before posting
every word in the transcript is linked to its second of audio. before a quote goes into a blog post or a podcast episode description, click the word — hear the guest say it. this catches the misheard names, the homophones, the dropped negation that auto-tools occasionally produce. it takes seconds per quote and prevents the publish-and-correct cycle.
private mode for confidential conversations
some meetings have NDAs attached. some have unpublished product details, financials, or partner names. some are just private by default. for those, run the file in private mode — audio stays on your laptop, transcript stays on your laptop, the export is local, no vendor in the chain. on a recent macbook (M2 or later) the local mode runs at roughly real-time on a one-hour file. same price as cloud mode. see /private-transcription/podcasting for the privacy posture in detail.
why we don't compete with the bot tools
if your job is twelve internal stand-ups a week and you want a summary in slack the moment each one ends, a bot tool is probably the right fit. that's not our shape. our user is a podcaster recording a guest, a blogger interviewing a source, an author recording a research conversation — someone whose meetings are content, not status updates, and whose subjects don't always want a third-party AI in the call. we sell to that user.
pricing for meeting transcription
$0.25 per minute. a 30-minute meeting is $7.50. a 60-minute meeting is $15. private mode and cloud mode are the same price. no subscription, no minimum, no per-seat fee. for podcast networks or content shops with steady volume, batch pricing arrives after launch — write hello@audiohighlight.com and tell us your shape.
waitlist signups get the lifetime deal: first month free, 50% off forever after.