What

Gemini Omni is a browser-based workspace for prototyping short AI-generated video drafts from text, image, video, and audio inputs, then revising them through natural-language editing. It appears aimed at teams and creators who need to test shot direction, reference consistency, pacing, and product communication before committing to formal production or full editing workflows.

Features

Multimodal video generation: Uses text, images, clips, and audio cues to create draft videos, which helps teams turn ideas and source assets into reviewable motion samples.
Conversational editing: Lets users describe changes in natural language so they can refine shots, subjects, style, and pacing without restarting from scratch each time.
Reference boundary control: Supports setting recognizable anchors such as people, products, layouts, or style constraints to reduce drift across remixes.
Test-oriented workflow structure: Organizes work around a shot question, reference lock, and cost/decision review, making each render easier to evaluate against a specific objective.
Credit and output planning: Shows generation choices such as duration, resolution, and estimated credit cost before submission, which helps teams prototype direction before spending more on longer or higher-quality runs.

Helpful Tips

Use this kind of tool for direction validation, not as an automatic replacement for final editing, captions, color finishing, or multi-track sound work.
Start with short, lower-risk renders to test one variable at a time, such as hook clarity, product detail, or subject consistency.
When identity or packaging matters, provide clear reference assets and explicit boundaries so review discussions focus on the intended decision rather than random variation.
Treat generated clips as drafts requiring rights, consent, brand, and platform review before publication; the page explicitly frames outputs as prototypes.

OpenClaw Skills

A likely OpenClaw fit is a workflow that turns Gemini Omni drafts into structured review and production decisions rather than standalone generation. For example, likely skills or agents could capture shot-test briefs, compare draft outputs against brand or campaign criteria, summarize next-version notes, route approved directions into editing or ad-production workflows, and maintain a searchable history of prompts, references, and review outcomes. If no native integration is stated, this should be treated as a plausible orchestration use case rather than a confirmed product capability.