Built for the full moment workflow

From long VOD to the moment you meant to keep.

ChatClipThat helps you find what mattered, judge candidates quickly, shape the edit, and publish only what is useful. Product teams can use the same video intelligence as timestamped JSON.

  • Multimodal moment discovery
  • Creator-controlled editing
  • Analyze before rendering
Find the signal

Search the recording, not just the transcript.

ChatClipThat combines available visual, transcript, audio, and chat context so discovery can follow what happened on screen as well as what someone said.

Semantic video search

Describe the moment you remember.

Search for a gunfight, helicopter, reaction, joke, or other concept and jump to matching timestamps without requiring an exact transcript phrase.

01:42:08Fast movement, low health, reaction spikeStrong match
02:19:31Related escape sequenceRelated
Voice trigger

Say "chat clip that."

Mark a live moment without stopping gameplay, then review it with the AI-discovered candidates after the VOD finishes.

Phrase detectedMarker at 01:42:19
AI-assisted discovery

More context than a volume spike.

Action, reactions, spoken context, available chat, and visual change help surface candidates worth a human decision.

  • Ranked candidate moments
  • Timestamped source context
  • Creator review before export
One continuous workflow

Discover. Decide. Deliver.

  1. 01

    Bring the footage

    Connect a supported VOD or upload authorized video.

  2. 02

    Find candidates

    Review AI suggestions, voice markers, or semantic searches.

  3. 03

    Choose the outcome

    Edit a clip, export it, or keep structured results as JSON.

Fast candidate review

Make the human decision without opening every timeline.

Swipe or tap through candidate moments. Keep the strong ones, skip weak ones, choose visibility, and open the editor only when a clip needs work.

  • Keep or skip quicklyReview the moment and its context before committing.
  • Private by default when neededChoose which clips belong on a public creator feed.
  • Edit with intentSend only selected candidates into the full editor.
AI candidate00:33THAT ACTUALLY WORKED
Last-second escapeAction + reaction + chat spike
94
SkipPrivateKeep
Creator-first editing

AI finds the candidate. You control the cut.

Turn selected moments into clips with deliberate hooks, layouts, overlays, and reusable style choices. Rendering begins only when the workflow asks for finished media.

01

Hook-first editing

Move the strongest beat to the opening, trim dead air, extend context, and keep the payoff intact.

02

Reframe and facecam

Choose vertical, square, or horizontal output and place gameplay and camera where the clip needs them.

03

Static text and captions

Add a clear setup, style readable captions, and keep text inside the format's safe area.

04

Chat overlays

Bring the audience reaction into the clip when chat adds context to the moment.

05

Reusable templates

Save caption, layout, text, and overlay choices so the next edit starts in your established style.

06

Montages and exports

Arrange selected beats into a montage or export individual clips for the destinations you support.

Keep the useful output

Clips for the feed. Evidence for the next decision.

Publish creator-facing work, keep clips private, download finished media, or leave the result structured for a product workflow.

Creator feed

One home for the moments you choose to share.

Present selected clips with your profile, banner, socials, and visibility controls instead of publishing every candidate automatically.

Export-ready media

Render the selected edit, not the whole analysis.

Download finished clips after review. Discovery and structured results do not force an MP4 render.

Developer API

Video intelligence without forcing a render.

Upload authorized recorded video, run reusable scanners, and receive typed observations with evidence timestamps and version provenance. Scanner runs are JSON-only; optional rendering is a separate one-off Moment workflow.

  • Saved monitor, summarizer, classifier, scorer, extractor, and find-moments scanners
  • One private upload can support multiple independently billed scanner runs
  • Explicit false answers remain distinct from unanswerable results
Follow the visual scanner quickstart
observation.jsonanswered
{
  "value": false,
  "value_type": "boolean",
  "semantics": "match_found",
  "start_ms": 0,
  "end_ms": 965000
}
Grounded resultFull-video evidence window
JSON only
typed answertimestampsno auto-render
Your next VOD already has moments in it

Find the ones worth keeping.

Start with the footage you already have, review the candidates, and stay in control of what becomes a clip.

Loading...
Downloads 0