Clips and boundaries

When an analysis finishes you have a list of clips. They are a starting point: you can manually edit them without a processing charge.

The list

Each analysis keeps its own clips. Open the Analysis # menu above the list to switch sets, or choose Re-analyze video… to try another set of directions. Switching saved results costs nothing; a new analysis creates new OpenAI usage on your own key.

Clips arrive strongest first. Each row can show a title, how long the clip runs, where it sits in the recording, a one-line description and a score - how strongly Obiklip rated it against the rest of the recording.

Sort the list by score, by length or by position in the recording. Use Columns to hide anything you do not want to look at; what you hide is remembered.

Click a row to select it. Clicking does not start playback unless you ask it to - the switch for that is next to the Columns button, and in Settings.

Playing one

A selected clip plays on the stage as its own little video: it starts where the clip starts and stops where it ends, with the subtitles drawn as they will be rendered. Nothing has been written to disk yet - this is a preview of the real thing.

Changing where a clip begins and ends

Open a clip to edit it. There are three ways to move a boundary:

Obiklip keeps a boundary on a pause in the speech rather than exactly where you let go, so a clip opens on a whole word instead of half of one. A little breathing room is left at each end on purpose: a clip that starts on the speaker's first syllable feels clipped.

Worked example: Fixing a cut that starts too early.

Titles and descriptions

The editor's Post Copy tab can also generate a hook, post title, caption, quote and hashtags. The request is billed to your own OpenAI account like any other. See Finish a clip for posting for the full workflow.

Every clip comes with a title written for it. Rewrite it - it is the file name when you export, and it is what you will scan the list for in a month.

The transcript

The transcript panel holds everything said in the recording. Search it to find a moment, click a line to send the video there, and correct any word that came out wrong. Corrections show up in the burned-in subtitles the next time you export.

The whole transcript can be downloaded as SRT subtitles or as plain text.

Trash

A clip you do not want goes to Trash, and the list gets shorter. Trash keeps it, so a clip thrown away in a hurry can be brought back; emptying Trash is what removes it for good. Neither action starts a new analysis.

Adding a clip Obiklip missed

Obiklip proposes; it does not decide. If there is a passage it walked past, select those lines in the transcript and press Clip. A new clip appears in the list, and behaves like any other from then on - trim it, title it, style it, export it. This is local editing and makes no new AI call.