Turn Riverside per-track recordings into LinkedIn video posts with Magic Clips, SRT captions and B2B-framed copy. Workflow plus Kompozy automation.
Published 2026-05-22 · Updated
Short answer
To repurpose a Riverside recording for LinkedIn, download the separate participant tracks, use Magic Clips as a shortlist and check the transcript for better moments, crop tight on a single speaker, reframe to square, attach an SRT caption file from desktop, and open the post with the takeaway. Riverside records up to 4K locally per participant, so tight crops hold up. LinkedIn accepts video up to 15 minutes and 5 GB.
Key takeaways
Riverside is one of the cleanest sources for this kind of work. It records each participant locally on their own device, in up to 4K video, and keeps every participant on a separate track, so each guest's file is as good as their camera and mic, whatever the call connection did. The workflow takes advantage of that: pick the best single-speaker track per clip, crop tight, and get cleaner clips than a typical call recording gives you.
Riverside also has Magic Clips, which pulls short highlight clips from the recording automatically. Even if you don't use Magic Clips directly, the per-track source gives your manual clipping more room to crop.
Podcasters and B2B executives who record on Riverside use this move to put their cleanest clips on LinkedIn. A high-resolution per-track source holds up to tight crops that would look soft from a standard call recording.
Riverside records up to 4K video and uncompressed 48 kHz audio locally, with each participant on a separate track. Magic Clips generates short clips from the recording automatically.
LinkedIn accepts native video up to 15 minutes and 5 GB, in aspect ratios from 1:2.4 to 2.4:1, so square and vertical cuts both fit. You can attach an SRT caption file when you post from desktop. Only the opening lines of a post show before "see more", so the hook goes first.
| Issue | Fix |
|---|---|
| Using a single mixed-down file when per-track 4K is available | Download the separate participant tracks. You get a cleaner file and more room to crop. |
| Trusting magic clips without checking | Treat magic clips as starting candidates; verify against the transcript for higher-density picks. |
| Burned captions only | Also attach an SRT file when posting from desktop. |
| Tagging the guest without consent | Confirm consent before social tagging. |
| Posting at random times | Schedule for the days and times your own LinkedIn posts do best. |
| Show-intro music in the clip | Trim the intro music; LinkedIn clips work better without music beds. |
A persona short, a vertical 9:16 video made in the Kompozy workspace. It shows the short-form format these clips end up in. It was not cut from a Riverside recording.
Best for:
Not ideal for:
Riverside records each participant locally, in up to 4K, on a separate track, so a weak connection doesn't degrade the file. Crops hold up better than on a standard call recording.
As starting candidates only. Verify against the transcript for higher-density moments the AI missed.
Per-track. Each speaker gets their own file, which gives you more room to crop and cleaner audio.
Make an SRT from the transcript and attach it on desktop, since LinkedIn needs an SRT file for closed captions. Burned-in captions are a good backup.
It depends on the episode. Pick every moment that stands on its own and stagger them over a couple of weeks.
Partly. You can upload a Riverside track to Kompozy, which transcribes it, scores moments, renders vertical 9:16 clips with burned-in captions and schedules them to LinkedIn. It does not read Riverside's Magic Clips, and square cuts need another editor.
We checked every platform figure on this page against the official sources listed below on 2026-10-23 and removed the ones we could not confirm, including Riverside recording length, clip counts and LinkedIn posting windows. Kompozy details come from kompozy.io/pricing and the Kompozy codebase.
Sources
Browse all repurposing pairs · See Kompozy pricing · Get Started →