3D Science Video Prompts for Google Flow | AI Prompt King

A familiar action makes an effective opening for a science explainer: a person lifts a cup of coffee, smells it, and the camera moves into a conceptual view of the nose. This guide splits that idea into three manageable scenes. Make each scene separately in Google Flow, then arrange the clips in an editor with your own narration.

The cover is an original AI concept illustration. It is not a verified frame from a generated Flow video. The prompts below are starting points; review the resulting motion and anatomy before publishing. This is a creative tutorial, not a medical diagnostic model.

What you need

• A Google account with access to Flow and enough credits for the selected model.
• A vertical project using a 9:16 aspect ratio.
• The three scene prompts below, submitted one at a time.
• A video editor for arranging clips, adding narration, captions and a quiet sound bed.

The AI Prompt King app is available through the APP LINK on @rahulgaur3d's Instagram profile or the Google Play link on this website. You can also find this guide by searching Google for AI Prompt King, opening aipromptking.in, and searching the website for 3D Science Video. The website route works without an app installation.

Step 1: Generate the coffee opening

Open Google Flow and create a project. Choose Video, select 9:16, and check the available model, duration and resolution settings. An eight-second scene is a useful starting point. Controls and credit costs can vary by model and account, so check the settings shown in your own session.

The page's main Copy Prompt button copies this first scene only. Paste it into Flow and generate one result. No private reference video or photo upload is required for this text-to-video example.

Scene 1 — Coffee aroma

Create an 8-second vertical 9:16 premium 3D science-explainer shot. An original fictional adult man in a plain teal T-shirt gently smells warm coffee from a plain navy mug in a deep-blue studio. Natural slow inhale, subtle steam, believable hands, stable face and mug, soft studio lighting. Slow camera push toward the nose. No text, logos, dialogue or music.

Keep the action restrained: one gentle inhale and a slow camera push are enough. Look for a stable mug, plausible fingers and a consistent face. If the subject spills the coffee, changes clothes or develops extra fingers, revise that issue and regenerate the scene before continuing.

Step 2: Show sensory tissue inside the nose

Create a second clip with the prompt below. This is a conceptual macro visualization, not a literal endoscope recording. The olfactory sensory tissue belongs high in the nasal cavity. The short sensory cilia should read as fine fibers in a thin mucus layer, not as long black hairs growing out of skin.

Scene 2 — Odor particles and sensory cells

Create an 8-second vertical 9:16 premium 3D educational science animation. Conceptual extreme macro view of olfactory sensory tissue high in the human nasal cavity. Show tiny gold odor particles drifting through air toward a thin clear mucus layer over softly shaded pink sensory epithelium. Delicate short translucent sensory cilia spread along the surface; brief subtle cyan receptor activation signals appear as particles interact. Slow controlled camera glide, shallow depth of field, stable organic geometry, soft warm light. Cilia are fine fibers, not black skin hairs or thick spikes. No blood, skin follicles, extra organs, labels, logos, narration or music. This is a conceptual scientific visualization, not a clinical scan.

The gold particles and cyan activity are visual conventions. They help a viewer follow the explanation; they do not represent the literal visible colors of individual odor molecules or nerve impulses. Avoid turning the scene into an unrelated tunnel of glowing spikes.

Step 3: Follow the signal toward the brain

Use a third clip to show a simplified side cutaway. Keep the nasal cavity, the sensory tissue high in its roof and the olfactory bulb in a plausible relationship. Fine olfactory fibers pass through the cribriform plate to the bulb; the olfactory tract carries information onward into the brain. Do not depict the spinal cord as the smell pathway.

Scene 3 — Olfactory bulb and brain

Create an 8-second vertical 9:16 premium 3D educational visualization in a deep blue studio. A clean simplified side cutaway of a human head shows the nasal cavity, olfactory sensory tissue high in the nasal roof, fine olfactory nerve fibers passing through the cribriform plate to the olfactory bulb above the nasal cavity, and the olfactory tract continuing toward brain regions involved in smell. Animate restrained small gold pulses from the high nasal sensory tissue to the olfactory bulb and along the tract toward the forebrain. Use biologically plausible spatial relationships, careful organic shading, a slow controlled lateral camera push and stable geometry. Do not route smell signals through the spinal cord or depict the olfactory bulb as the cerebellum. No gore, readable labels, logos, dialogue or music. Conceptual educational visualization; avoid excessive glowing neural decoration.

A generated anatomy shot can look polished while still containing mistakes. Compare the structure and direction of the pulses with a reliable anatomy reference. If the model repeatedly fails the cutaway, simplify the view and reduce the number of moving elements. Do not label an incorrect result as an accurate anatomical demonstration.

Step 4: Download and assemble the clips

After generation finishes, inspect each clip and use the download control on the result. Put the coffee scene first, sensory tissue second and the brain pathway third. Trim each shot to the useful action; you do not need to use every generated second. Match the deep-blue palette and avoid a major character or lighting change across related shots.

Add original Hindi narration in the editor after arranging the visual beats. Keeping narration separate makes pronunciation repairs easier and avoids relying on generated scene audio. Captions should follow the spoken meaning, remain readable on a phone and stay away from interface overlays. Use short cuts or gentle transitions. Repeating a loud whoosh on every edit usually competes with the explanation.

Example Hindi science narration

कॉफी की खुशबू के अणु नाक के ऊपरी हिस्से तक पहुँचते हैं। वहाँ मौजूद गंध पहचानने वाली कोशिकाएँ इनसे सक्रिय होती हैं। अलग-अलग खुशबुओं से संकेतों के अलग पैटर्न बनते हैं। ये संकेत मस्तिष्क तक पहुँचते हैं, जिससे हमें खुशबू पहचानने में मदद मिलती है।

This optional narration explains the science example. It is separate from the Rahulgaur3D tutorial voice that explains how to use Flow. Use your own recording or an authorized synthetic voice, and check the Hindi pronunciation before export.

Export checklist

• Keep the final frame vertical, normally 1080 × 1920 for a finished Reel.
• Confirm that the source preview, generated example and editing instructions are clearly distinguishable.
• Check the face, hands, coffee mug, cilia and signal direction at normal playback speed.
• Keep narration intelligible and background audio quiet underneath it.
• End immediately after the final spoken line; remove accidental empty frames.
• Watch the exported MP4 on a phone before posting.

Common questions

Can I paste all three prompts at once? Generate them separately. Each prompt describes a different shot. Combining them in one short request may cause rushed transitions, missing actions or a scene that never reaches the brain view.

Do I have to upload the reference creator's face? No. These prompts create an original fictional character from text. If you later need the same character in multiple shots, use a supported reference workflow with an image you are allowed to use. Do not assume that writing the same description guarantees an identical face.

Is a specific model or price guaranteed? No. Model names, resolutions, durations and credit costs depend on current availability in your Flow account. Check the generation panel before submitting. This guide does not promise free or unlimited generation.

Will the generated anatomy always be correct? No. Prompt instructions guide the model but do not verify the output. Check educational scenes against reliable sources and regenerate incorrect shots.

Sources and further reading

• Google Flow Help: https://support.google.com/flow/answer/16353334?hl=en
• NIDCD, Journey of Smell to the Brain: https://www.nidcd.nih.gov/news/multimedia/journey-of-smell-to-brain-text-version

Guide prepared for AI Prompt King and @rahulgaur3d. Updated 26 October 2026.