Gemini Omni Prompt Guide: 10 Copy-Paste Prompts for Multimodal Video Generation
At Google I/O 2026, Google unveiled Gemini Omni with one promise: “create anything from any input” — starting with video. The first model in the family, Gemini Omni Flash, went live on May 19, 2026: one multimodal prompt in (text, images, audio, video clips — any combination), one 10-second video out, with synchronized audio and chat-based editing. This is not Veo with a new name — it’s a different prompting workflow. Below: what the model actually does (checked against Google’s launch materials and API docs), how to structure prompts for it, and 10 copy-paste templates.
What Gemini Omni Actually Does
Verified capabilities as of this writing — and one honest caveat: this surface is moving fast, so re-check Google’s docs before building a pipeline on it.
- True multimodal input. Text + image + audio + video can go into a single instruction, and the model reasons across them rather than stitching outputs together. Hand it a character photo, a dialogue track and a lighting reference clip, and one output resolves all three constraints.
- 10-second clips with synchronized audio. Google has said the 10-second cap is a product decision, not an architectural limit — so expect it to move.
- Conversational editing. “Make the lighting warmer” edits the current clip in chat instead of regenerating from scratch — the biggest workflow shift versus prompt-and-regenerate models.
- Reference workflows. The developer preview (
gemini-omni-flash-previewvia the Gemini API Interactions interface) exposes text-to-video and reference-to-video task types, with reference images guiding the output. - SynthID watermarking embedded in every generated clip — not optional.
- Where it runs: the Gemini app, Google Flow, YouTube Shorts and YouTube Create, with paid access through Google AI plans (starting at Google AI Plus, $7.99/mo). A higher-end Omni Pro is announced but has no release date.
- What’s deliberately missing: editing speech or audio inside generated clips is being withheld for now. Don’t design production workflows that depend on it.
How to Prompt Gemini Omni
Gemini Omni thinks in constraints: content, format, references and sound are all inputs it weighs together. Write prompts the same way.
[Task: text-to-video | reference-to-video] [shot type] of [subject + material],
[beats], [environment + time of day], [camera move], [lighting], [style/grade],
[mood], [format: aspect ratio + duration], [audio: voice-over / SFX / ambience].
Three rules for multimodal prompts:
- Put format and duration in the prompt. “9:16, ~7 seconds” is treated as a constraint to reason over — inline format notes measurably cut re-rolls.
- References beat adjectives. A 30-word prompt with three solid references (character photo, style frame, motion clip) outperforms a 300-word prompt with none — just say what each reference contributes.
- Use audio as an input, not just an output. Synced audio is native — write the exact voice-over line or sound cue into the prompt. Just don’t expect to re-voice a finished clip later; that edit path is withheld.
10 Copy-Paste Gemini Omni Prompts
Natural Wonders
1. Aurora over the fjord
Wide aerial shot of green and violet aurora rippling over a dark Norwegian fjord,
stars reflected in still black water, camera gliding slowly forward above the
shoreline, cold silver-blue moonlight, documentary realism, awe-struck mood,
audio: soft wind and a single low synth pad. 16:9, 8 seconds.
2. Monsoon waterfall
Drone shot circling a jungle waterfall in heavy monsoon rain, mist exploding off
wet black basalt, saturated green foliage bending in the wind, smooth orbit left,
diffuse overcast light with one break of sun, epic natural-power mood, audio: roar
of falling water and rain on leaves. 16:9, 10 seconds.
Urban Stories
3. The noodle vendor
Medium shot of an elderly noodle vendor in a Bangkok night market, steam rising
from a wok, warm tungsten bulbs and neon signs glowing on wet ground, he hands a
steaming bowl toward camera and smiles, slow push-in, lively documentary mood,
audio: sizzling wok, crowd chatter, ladle clangs. 9:16, 8 seconds.
4. Subway violinist
Medium shot of a violinist playing in a tiled subway station, commuters blurring
past in motion, cool fluorescent light with a warm tunnel glow at the far end,
camera slowly circling her, lonely beautiful mood, audio: solo violin melody over
low train rumble. 9:16, 10 seconds.
Product Macro
5. Perfume hero macro
Macro product shot of a faceted glass perfume bottle on black stone, a single drop
of amber oil rolling down one facet, caustic light patterns rippling on the stone,
slow dolly-in, warm 3200K key light with deep shadows, luxury commercial grade,
audio: low atmospheric drone and one soft bell strike at the end. 1:1, 7 seconds.
6. Espresso pour
Extreme macro of espresso pouring into a clear glass, crema swirling in golden
ribbons, steam curling through a hard side light, slow vertical tilt following the
stream, dark warm palette, tactile satisfying mood, audio: rich pour and faint
grinder hum. 9:16, 6 seconds.
Sci-Fi Concepts
7. Terraforming drones
Wide shot of a swarm of insect-like terraforming drones seeding moss over red
Martian rock at dawn, dust plumes drifting through low orange sunlight, one drone
lands close to camera and unfolds its legs, slow forward tracking shot, grounded
hard-sci-fi realism, audio: servo whirrs and thin wind. 16:9, 10 seconds.
8. The memory archive
A researcher walks through a dark archive hall where holographic memory scenes
float inside glass columns — a birthday party, a city street in the rain — casting
cyan light on her face as she passes, slow lateral track, cold blue palette with
warm hologram accents, melancholy sci-fi mood, audio: soft hum and faint echoes of
the memories' sound. 16:9, 8 seconds.
Documentary Texture
9. Claymation explainer with voice-over
Stop-motion claymation style: a glossy clay protein chain folds itself into a
helix on a workbench of clay tools, camera gently pushing in, warm tabletop
lighting, visible fingerprints in the clay, educational playful mood, audio: calm
voice-over saying "Proteins fold into precise shapes — and the shape is the
function." 16:9, 10 seconds.
2D Animation
10. The growing city
Flat 2D vector animation: a small robot character waters a potted plant that grows
into a full city skyline in time-lapse, clean geometric shapes, bold limited
palette of teal, coral and cream, subtle paper texture, static medium shot,
optimistic storybook mood, audio: light marimba motif and a soft watering-can
pour. 16:9, 8 seconds.
The Multimodal Playbook: References and Context
This is where Omni changes the game — references carry more signal per byte than prose:
- Character lock — attach a portrait, write “the person in the attached photo,” and describe only the action. Keeps identity stable across clips.
- Style lock — attach a frame from a look you like: “match the grade and composition of the attached frame.”
- Motion lock — attach a short clip: “use the camera move from the attached reference clip.”
- Beat lock — attach a music track: “cut the visual beats to the attached track.” Made for Reels and music edits.
- Context line — when several references ride together, label them in the prompt: “[face ref] for the character, [frame ref] for the grade.” Ambiguous references get blended into mush.
- Edit in chat — don’t regenerate for one fix. Follow up with “keep everything, make the lighting warmer” or “swap the dog for a cat.” On later turns, restate the one thing that must not change (“keep the character and camera move identical”).
FAQ
Is Gemini Omni the same as Veo? No. Veo 3.1 is text/image-to-video with synchronized audio and prompt-based re-generation. Gemini Omni accepts text, image, audio and video in one instruction and edits clips conversationally in chat. Different architecture, different workflow.
How long can a Gemini Omni video be? 10 seconds per generation at launch — Google explicitly framed the cap as a product decision rather than a model limitation, so it may be relaxed over time.
Can I use reference images? Yes. Reference-driven generation (the API’s reference-to-video mode) is a core use case: character photos, product shots, style frames and short motion clips can all steer the output. Describe what each reference contributes.
Does it generate audio? Yes — synchronized audio ships with the video, and input audio can act as a generation constraint (a dialogue line, a music beat). What’s withheld is editing speech or audio inside already-generated clips.
Is there an API?
A developer preview (gemini-omni-flash-preview) is available through the Gemini API’s Interactions interface; consumer access runs through the Gemini app, Google Flow and YouTube’s creation tools. Check Google’s official docs for the current access and limits before you automate anything.
Multimodal prompts reward iteration: set up your references, copy a template, tweak in chat until it clicks. Browse what other creators are shipping at VideosPrompt — and when the clips are ready to go out, a tool like FLOWNIB can push the same cut across multiple platforms in one pass.
Share Article
