what we make for you, after the cut
a clip pack you can actually post
when a guest says something quotable, you mark it. we hand back a list of clips — each with a verified timestamp, the surrounding context, and the cleaned-up text — ready to drop into a social post or a YouTube short. the timestamp is the point: every clip you post can be traced back to the second of audio, so when a fan or a critic asks "did they really say that?" you've got the receipt.
corrected names, every episode
custom-vocabulary loaded once, applied forever. your guest's name spelled right. their book title spelled right. the jargon of your beat — horse racing, biotech, fantasy football, whatever — spelled right. this is the feature incumbents used to have and have been quietly removing; it's our differentiator on purpose. load your beat's rolodex, never re-fix the same surname twice.
a blog draft, not a wall of timestamps
the transcript exports your editor produces are technically correct and practically unusable as the basis for a blog post. we hand back a markdown draft that reads like prose: paragraph breaks where speakers change, filler trimmed, attribution lines preserved, key quotes pulled to the top with a bold lede. you edit it for ten minutes; you don't rewrite it from scratch.
see the blog-post export format page for what the output actually looks like.
a searchable archive across years
every episode you've ever recorded, queryable by transcript. "every time we mentioned the Fed in the last two years." "every time my co-host said 'I disagree.'" "every guest who mentioned a specific bill number." most podcast platforms treat the transcript as a per-episode artifact and forget about it; we treat it as a corpus you own.
workflow
- record in whatever you already use. logic, riverside, descript, zoom, a field recorder, a phone — if it produces a file, we work with it.
- edit the audio in your editor of choice. when the cut is locked, drop the final mix into audiohighlight.
- review the transcript. fix the speaker labels for "host" and "guest" once each — they propagate. add any names or jargon to your rolodex if the model missed them.
- mark the quotable moments as you read. those become the clip pack. mark the section breaks; those become the blog draft's paragraph structure.
- export the artifacts you actually need. show notes for the podcast page, blog draft for substack/medium/your CMS, clip pack for social, captions for the video version, full .docx for the editorial archive.
when private mode matters
most podcast transcription is fine on a cloud tool. you record a public conversation, you transcribe it, you post the show. there's nothing to protect.
but some episodes have things that haven't aired yet — an embargoed product launch your guest mentioned, an off-the-record passage you're trimming before publishing, a guest who agreed to come on under the condition that their audio doesn't go to additional third parties. for any of those, switch to private mode. transcription runs in your browser, on your machine; the audio never makes a network request. when you close the tab, the transcript is yours and ours is gone.