The worst part of making highlights isn't the editing — it's the hunting. You sit through a three-hour VOD dragging the playhead, terrified you'll miss the one moment that mattered. Obskura flips that around: it finds the candidate moments for you, and you just decide what's worth keeping.

Let the recording analyze itself

After Obskura captures a stream, it analyzes the footage on-device and scores it second by second across a few signals: audio loudness, on-screen motion, and scene changes. Big spikes in those signals are where interesting things tend to happen — a teamfight, a clutch, a sudden reaction. You don't tag anything; the analysis runs automatically.

Get a shortlist of clip candidates

From those signals, Obskura produces clip candidates — suggested start/end ranges where something noteworthy likely occurred. Each candidate comes with a source label (Highlight, Speech, Motion, or Scene) and a confidence score, so you can see why a moment was flagged. Instead of three hours of video, you're now looking at a short list of "here are the moments worth your attention."

Keep, reject, trim, export

Candidates land in a timeline-style editor where the workflow is deliberately simple:

1. Review each suggested clip — keep the good ones, reject the rest.
2. Nudge the in/out points to tighten the trim if you want.
3. Export the keepers as clean files, ready to post or drop into your editor.

Because the heavy lifting (finding moments) is already done, a highlight pass that used to take an evening becomes a few minutes of keep/reject. We covered the release of this clip-finding workflow in more detail in our clip search announcement.

A realistic example

Say you streamed a four-hour session. Obskura's analysis flags 18 candidates — loud moments, big motion swings, scene cuts, and spots where you reacted vocally. You skim the 18, keep 6, trim two of them, and export. The four hours of raw VOD never needed a single manual scrub.

Why it stays on your machine

All of this analysis happens locally. Your footage and audio aren't uploaded to a cloud service to be processed — the scoring, the candidate detection, and the export all run on your own hardware. That keeps it fast and keeps your content yours.

If clipping highlights is the main reason you record, this is the core of what Obskura does. See the full feature set, or check how it compares to other livestream recorders.