Welcome to the Gemini Omni API and Prompts repository! 🤗 We collect high-quality prompts and video examples for Google Gemini Omni across a wide range of creative tasks including transform, motion, camera control, text sequences, and multi-input workflows. Most cases in this repository are curated from DeepMind official demos, prompt guides, and community experiments. Try it on Free Gemini Omni: Gemini Omni If you find this useful, consider giving it a star. ⭐
Note
This repository focuses on reusable prompt patterns and reference cases for Gemini Omni video generation on Free Gemini Omni.
- May 25, 2026: Added 16 new community cases + new Community Gallery section
- May 23, 2026: Added 10 community showcase cases from trending tweets
- May 22, 2026: First repository update with 25 curated Gemini Omni prompts
- 🎯 Prompt Ingredients
- ✂️ Edit
- 🔄 Element Replacement
- Case 1: Butterfly to Bee
- Case 2: Bee to Fireflies
- Case 3-5: Spaceships & Astronaut Series
- Case 6: Train 1896 Transform (by @emollick)
- Case 7: Remove Person from Video (by @arrakis_ai)
- Case 8: Invisible Violin
- Case 9: Location Change via World Knowledge (by @venturetwins)
- Case 10: Animation to Live Action (by @arrakis_ai)
- 🎬 Base Scene
- 📷 Camera Direction
- 🎬 Action & Sync
- 🔄 Element Replacement
- 🎨 Advanced Multi-Modal
- 🪞 Artistic Style
- ✨ Visual Effects
- 🔗 Cross-Modal
- Case 1: Transport to New Environment
- Case 2: Birds Shape With Audio
- Case 3: Slide to Motion (by @yoshifujidesign)
- Case 4: Isometric Cooking Character with Reference Image (by @kumiko_shiraki)
- Case 5: ChatGPT Instruction Image as Input (by @Majin_AppSheet)
- Case 6: ChatGPT Illustration to Omni Animation (by @mmmiyama_D)
- 📋 Storyboard
- 🔤 Text Rendering
- ⚖️ Comparison
- 🧪 Evaluation
- 🌐 Community Gallery
- 🙏 Acknowledge
Gemini Omni has strong world understanding — it draws on real-world knowledge of history, science, and culture. You don't need to over-explain every detail. Instead, express your creative intent in natural language and let Omni's reasoning fill in the rest. When creating a new video from scratch, mix these dimensions to control the output:
| Dimension | What to specify | Example |
|---|---|---|
| Shot Framing & Motion | Wide-angle, medium, or close-up. Camera trajectory: gentle glide, sudden rush, static lock, dolly zoom, etc. | A close-up tracking shot smoothly pushing in |
| Style | Overall visual art direction | Vintage monochrome hologram, 3D voxel art, colored crayon aesthetic |
| Lighting | Scene mood and light setup | Warm champagne lighting, dim overhead gym lights |
| Location | Environment and background | Small underground gym, futuristic neon cityscape |
| Action | Subject behavior and movement | The person touches the mirror, a marble rolling fast on a chain reaction track |
Tip
Iterative Editing: Omni supports multi-turn conversation editing. It preserves what works and only modifies what you ask — no need to re-describe the entire scene each time. Just say what to change next.
Tip
Preserve Unchanged Areas (by @tanabe_fragm): When editing video, add phrases like "Don't change anything else" or "Keep everything else the same" to your prompt. This significantly reduces unwanted changes to parts of the video you didn't intend to modify.
045_tanabe_fragm_preserve_prompt_tip.mp4
|
Input: input-gemini-omni__prompt-guide__butterfly.webm |
Output: 015_butterfly_to_bee.webm |
Prompt:
Change the butterfly to a bee.
|
Input: 015_butterfly_to_bee.webm |
Output: 016_bee_to_fireflies.webm |
Prompt:
Change the bee into a small swarm of fireflies.
|
Input: input-gemini-omni__prompt-guide_astro-spaceship.webm |
||
023_spaceships_white_origami.webm |
024_astronaut_sea_anemone.webm |
025_small_ships_to_stingrays.webm |
Prompts:
Case 3: Change the ships to be made from white origami paper.
Case 4: Change the astronaut to a sea anemone.
Case 5: Change the small ships to stingrays.
Case 6: Train 1896 Transform (by @emollick) 🎬 Video→Video
036_emollick_train_transform.mp4
Prompt:
I took the famous "train" movie from 1896 & made it a bullet train, LEGO, added a time traveler, a centipede, muppets...
Case 7: Remove Person from Video (by @arrakis_ai) 🎬 Video→Video
038_arrakis_ai_remove_person.mp4
Prompt:
Remove the person from this video perfectly.
|
Input: 010_violinist_transport_environment.webm |
Output: 011_violinist_invisible_violin.webm |
Prompt:
Make the violin invisible
Case 9: Location Change via World Knowledge (by @venturetwins) 🎬 Video→Video
047_venturetwins_location_change.mp4
Prompt:
Re-shoot this video in [location] based on the screenshot from Google Maps.
Uploaded a Waymo ride video, then asked Omni to re-shoot in different locations using Google Maps screenshots. The model leverages its world knowledge to seamlessly change environments.
Case 10: Animation to Live Action (by @arrakis_ai) 🎬 Video→Video
051_arrakis_ai_animation_to_live.mp4
Prompt:
Turn this animation into live action.
009_violinist_input.webm
Prompt:
A video of a violinist playing a song.
|
Input: 011_violinist_invisible_violin.webm |
Output: 012_violinist_over_shoulder.webm |
Prompt:
Change the camera angle to be over the violinist's shoulder.
|
Input: input-gemini-omni__prompt-guide__butterfly.webm |
Output: 020_camera_tilt_boots.webm |
Prompt:
Change the camera angle, a close-up on his shoes, quickly tilting up to medium shot, then widening.
Case 3: Travel Selfie Hyperlapse (by @ZaraIrahh) 🔤 Text→Video
028_travel_selfie_hyperlapse.mp4
Prompt:
Create a 10-second cinematic hyper-lapse selfie travel video of the uploaded female character across 20 world-famous destinations in 2026. Hard cuts every 0.5 seconds synced to the beat. Handheld selfie-stick camera, wide-angle lens, close selfie framing, energetic travel-vlogger style, vibrant cinematic colors, realistic lighting, dynamic motion blur, natural crowds, and clear landmarks in every shot.
Case 4: Fashion Drone Shot (by @ariaxawan) 🔤 Text→Video
030_fashion_drone_shot.mp4
Prompt:
A 10 second ultra cinematic hyper realistic FPV fashion drone shot filmed in a single continuous take inside a futuristic luxury tunnel. Single continuous take, aggressive FPV motion, ultra smooth cinematic flight path, luxury high-fashion editorial atmosphere.
Case 5: Top View to 360 Rotation (by @npaka123) 🖼️ Image→Video
040_npaka123_360_rotation.mp4
Prompt:
この教室の中央から黒板を見ているファーストパーソンなゲーム視点。360度カメラを回転。教室の黒板右側の窓の外は廊下、黒板左側の窓の外は校庭。
Case 6: Omnizoom — Diving Into a Photo (by @alexanderchen) 🖼️ Image→Video
046_alexanderchen_omnizoom.mp4
Prompt:
Omnizoom — diving into a photo.
007_animal_toy_sound.webm
Prompt:
When the finger in <video> touches the animal toy play the sound the animal makes
|
Input: input-gemini-omni__prompt-guide__building-no-light.webm |
Output: 008_apartments_lights_sync.webm |
Prompt:
The lights of the apartments start turning on in sync with the music.
013_marble_chain_reaction.webm
Prompt:
A marble rolling fast on a chain reaction style track, continuous smooth shot
|
Input: input-gemini-omni__prompt-guide__building-no-light.webm |
Output: 017_prompt_guide_building_lights.webm |
Prompt:
The lights of the buildings start turning on in sync with the music.
Case 5: Boxing Fight Realistic (by @RuzainaMeer) 🔤 Text→Video
031_boxing_fight_realistic.mp4
Prompt:
Ultra-realistic 10-second boxing fight between two women inside a small underground gym. Both fighters look naturally athletic with realistic skin texture, sweat, bruises, and detailed facial expressions. The fight feels raw and authentic, like real professional sparring footage. The camera moves handheld around the ring at close range, capturing fast punches, defensive movement, realistic footwork, and heavy breathing.
001_mirror_liquid_metal.webm |
002_mirror_line_art.webm |
003_mirror_puppet.webm |
Prompts:
Case 1: When the person touches the mirror, make the mirror ripple beautifully like liquid, and the person's arm turns into reflective mirror material
Case 2: When the person touches the mirror, the person transforms into a detailed monochrome line art drawing
Case 3: When the person touches the mirror, the person suddenly transforms into a cute felted stuffed puppet version with large googley eyes and glasses
Case 4: Animation Ads One Shot (by @DenneyDara) 🔤 Text→Video
027_animation_ads_one_shot.mp4
Prompt:
Make a Pixar-style video of an aloe leaf that is walking through the forest that talks about how good nature makes it feel. Have it say, "Organic and healthy ingredients make me feel so good."
Case 5: Line Drawing Isolation (by @alexanderchen) 🎬 Video→Video
029_line_drawing_violin.mp4
Prompt:
Extract the key object in this video. Render a video showing that object as a black diagram-style line art drawing on solid 100% white background, nothing else in background. Keep the motion and sound exactly as is.
006_handhole_super_zoom.webm
Prompt:
Make it look like the weird shape of my hand hole super zooms and magnifies the ground it's looking at in sharper quality.
019_skateboard_motion_effects.webm
Prompt:
Edit this keeping everything the same. Add animated motion effects coming out of the skateboard.
Case 3: AR HUD Overlay (by @jerrod_lew) 🎬 Video→Video
052_jerrod_lew_ar_hud.mp4
Prompt:
Create a virtual HUD and UI overlay for this recorded phone video, like an AR glasses experience with secondary screens.
|
Input: 009_violinist_input.webm |
Output: 010_violinist_transport_environment.webm |
Prompt:
Transport the violinist to the image environment
|
Input Video: input-shape.webmInput Image: 
Input Audio: input-audio.webm |
Output: 021_birds_shape_with_audio.webm |
Prompt:
The birds from <video> loosely form the imperfect shape of a bird based on <image>. They move to the music from <audio> and dissipate as they fly
Case 3: Slide to Motion (by @yoshifujidesign) 🖼️ Image→Video
041_yoshifujidesign_slide_motion.mp4
Prompt:
GPT image2でスライド作成 → Gemini Omniでモーション。画面遷移もさせられるし、イラストの動かし方も自然。
Case 4: Isometric Cooking Character with Reference Image (by @kumiko_shiraki) 🖼️ Image→Video
048_kumiko_shiraki_isometric_cooking.mp4
Prompt:
Narrow down reference images and add negative prompts to get closer to your ideal output.
Technique: When the generated video doesn't match your vision, (1) narrow down reference images, and (2) add negative prompts to suppress unwanted elements.
Case 5: ChatGPT Instruction Image as Input (by @Majin_AppSheet) 🖼️ Image→Video
|
Input (instruction images from ChatGPT): 
|
Output: 049_majin_appsheet_instruction_image.mp4 |
Workflow: Generate instruction/storyboard images in ChatGPT, then feed them directly to Gemini Omni as visual prompts.
Case 6: ChatGPT Illustration to Omni Animation (by @mmmiyama_D) 🖼️ Image→Video
050_mmmiyama_d_chatgpt_to_omni_0.mp4 |
050_mmmiyama_d_chatgpt_to_omni_1.mp4 |
050_mmmiyama_d_chatgpt_to_omni_2.mp4 |
Workflow: Generate illustration diagrams with ChatGPT image generation → animate them with Gemini Omni. Text rendering can be improved by adding specific prompts to suppress text corruption.
Case 1: Luxury Cosmetic Commercial (by @aiwithaly) 🔤 Text→Video
032_luxury_cosmetic_commercial.mp4
Prompt:
Create a cinematic 10-second ultra-realistic luxury cosmetic commercial in a high-end skincare advertisement style. Use warm champagne lighting, glossy beauty-film aesthetic, shallow depth of field, macro beauty cinematography, smooth cinematic camera movement. 10 scenes from macro serum droplets to final payoff shot.
|
Input: 
|
Output: 022_storyboard_show_me.webm |
Prompt:
Show me in this story. Follow the story exactly in order starting top left. Entire story in 10 seconds. Cinematic
Case 3: 3x3 Split Screen (by @alexanderchen) 🎬 Video→Video
033_cinematic_story_cyberpunk.mp4
Prompt:
Generate a 3x3 split screen video based on different details you see here. Make each cell different, varying the perspective, composition, zoom, angle, camera movement (some static, some moving). Make some of the cells extreme close-ups with detailed textures. Keep it photorealistic, handheld, raw. Only natural sounds.
Case 4: Action Replay from Different Angles (by @jerrod_lew) 🎬 Video→Video
037_jerrod_lew_action_replay.mp4
Prompt:
Gemini Omni can create action replays from different angles. I referenced a video clip with agent instructions to generate replays.
Case 5: Split Screen Video (by @jerrod_lew) 🎬 Video→Video
039_jerrod_lew_split_screen.mp4
Prompt:
Use a reference video and ask the agent for a split screen video.
014_alphabet_items_sequence.webm
Prompt:
The video shows items of the alphabet. An unusual item starting with each letter is shown sitting on a table. All 26 letters must be represented by 26 items with matching lower thirds displaying the letter. Only one item and lower third at a time. Rapid fire, roughly 9 frames per item at 24FPS. Last frame is a slip of paper "THE END".
018_word_by_word_text_sync.webm
Prompt:
word by word, one word on the screen at a time: did, you, know, that, this, model, can, do, pretty, good, text!? each word appears with a different animated style, perfect pacing to a rhythm, sizzle reel.
Case 3: Text Rendering AI News (by @chrisfirst) 🔤 Text→Video
026_text_rendering_ai_news.mp4
Prompt:
Static shot we see them turn the page 3 times. Every flip we see content on both left and right side of book pages. Each contains a big news story around AI for the year of 2025. Include images and crystal clear text.
Case 4: Font Fashion Show (by @HBCoop_) 🔤 Text→Video
035_font_fashion_show.mp4
Prompt:
Create a 10-second avant-garde fashion editorial where every outfit is inspired by a specific Google Font personality. Each second introduces a new model styled around fonts like Playfair Display, Bebas Neue, Orbitron, Pacifico, Rubik Mono One, and Cormorant Garamond. Font names appear integrated into the environment using their exact typography style. High-fashion runway cinematography with bold lighting, mirrored sets, and surreal motion.
Case 1: Seedance 2.0 vs Gemini Omni Flash (by @JSFILMZ0412) ⚖️ Comparison
042_JSFILMZ0412_seedance_vs_omni.mp4
Seedance 2.0 Fast vs Gemini Omni Flash — style transfer, motion quality, and censorship comparison.
Case 2: Gemini Omni vs Seedance 2.0 Action Scenes (by @CuriousRefuge) ⚖️ Comparison
043_CuriousRefuge_omni_vs_seedance.mp4
Gemini Omni video editing model vs Seedance 2.0 — big action scenes comparison.
Case 1: Gemini Omni Quality Evaluation (by @kenichiota0711) 🧪 Evaluation
044_kenichiota0711_quality_eval.mp4
ほぼ理想的なアウトプットで感動。わずかに揺れる花と髪、まばたきの表現、ノートに書く動作の滑らかさ。
Creative experiments and showcases from the community. These cases demonstrate the breadth of what's possible with Gemini Omni.
|
Concept-Driven Educational Video — by @VORTEX_Promos 053_vortex_promos_concept_driven.mp4 |
054_paji_a_showcase.mp4 |
|
Nano Banana for Video — Consistency Test — by @WolfRiccardo 055_wolfriccardo_nano_banana.mp4 |
Presenting Isometric Character — by @kumiko_shiraki 056_kumiko_shiraki_presenting.mp4 |
|
One-Sentence Cinematic Zen — by @Dheepanratnam 057_dheepanratnam_cinematic_zen.mp4 |
Character Transformation — Glamorous to Everyday — by @HBCoop_ 058_hbcoop_character_change.mp4 |
|
Waymo to India — by @iHarnoorSingh 059_iharnoorsingh_waymo_india.mp4 |
The Letter Never Sent — Short Film — by @Strength04_X 060_strength04_x_letter_never_sent.mp4 |
This repository was inspired by excellent open prompt collections and community-shared examples.
Thanks to Google DeepMind for publishing official Gemini Omni demos and prompt guides that made these case studies possible.
Community Contributors:
@emollick, @jerrod_lew, @arrakis_ai, @npaka123, @yoshifujidesign, @chrisfirst, @DenneyDara, @ZaraIrahh, @alexanderchen, @ariaxawan, @RuzainaMeer, @aiwithaly, @HBCoop_, @JSFILMZ0412, @CuriousRefuge, @kenichiota0711, @tanabe_fragm, @venturetwins, @kumiko_shiraki, @Majin_AppSheet, @mmmiyama_D, @VORTEX_Promos, @paji_a, @WolfRiccardo, @Dheepanratnam, @iHarnoorSingh, @Strength04_X
If anything needs to be corrected, please contact us and we will update it.
If you have more interesting prompt cases to share, feel free to reach out and help us expand the Free Gemini Omni prompt library.