Learn

Prompt
Guide

12 min read All plan levels
The better your prompt, the better your storyboard. This guide teaches you exactly how to communicate with GoroSolution AI — from a one-line idea to a full cinematic scene breakdown. No prior AI experience needed.

01 The Basics — How the AI Reads Your Input

GoroSolution uses your text to do two things simultaneously: write the story (scene descriptions, story beats, act structure) and generate visuals (image prompts for AI rendering). These are separate processes — and understanding this makes a huge difference.

  • Film Concept → AI reads this to understand your story, characters, world, and tone. This feeds scene generation.
  • Image Prompt (inside Edit Scene) → AI reads this to generate the actual visual frame. Needs to be visual and specific.
  • Video Prompt → Describes a full 8-second motion sequence for AI video engines like Veo3.

Think of it like this: the Film Concept is your pitch to the director. The Image Prompt is your instruction to the cinematographer.

02 Writing Your Film Concept

The Film Concept box is the most important input. This is where you tell the AI what your story is about. You can write anything from a single sentence to thousands of words — the AI handles both.

One-Line Concepts (Fast)

Great for quick storyboards and exploration. Be specific about genre, setting, and character:

Too vague
A man goes on a journey.
Much better
A disgraced samurai in 1600s Japan seeks redemption by protecting a village from a warlord's army — filmed in dark, rain-soaked cinematography with long silences and sudden violence.

What to Include in Your Concept

  • Who — Describe your main character with specifics: age, appearance, emotional state, backstory hint
  • What — The core conflict or goal driving the story
  • Where — Setting, time period, world details (city, rural, dystopian, fantasy)
  • Tone — Dark, hopeful, comedic, epic, quiet, tense
  • Visual reference — "Filmed like Blade Runner 2049" or "documentary handheld style" helps enormously

Long-Form Concepts (Best Results)

Paste a full screenplay treatment, character sheets, world-building notes — GoroSolution handles up to thousands of words. The more context you give, the more consistent and accurate your scenes will be. Characters stay visually consistent. Settings match across scenes. Tone remains coherent.

Pro tip: If you have an existing script, paste the relevant scenes. The AI will extract the visual essence from prose automatically.

03 Writing the Image Prompt

The Image Prompt is your direct instruction to the image generation AI. It controls exactly what appears in the frame. GoroSolution pre-fills this from your concept, but you can edit it for precise control.

The Formula That Works

Follow this order in your image prompt for best results:

  1. Subject — Who or what is the main focus. Include full character description.
  2. Action — What they are doing right now in this frame
  3. Setting — Environment, location, time of day, weather
  4. Lighting — This is the single most powerful visual modifier
  5. Camera — Angle and distance (close-up, wide shot, low angle, bird's eye)
  6. Colour & tone — Colour grade, saturation, temperature
  7. Style — Any aesthetic reference (cinematic, oil painting, etc.)
Weak image prompt
A woman standing in the rain looking sad.
Strong image prompt
Maya (28yo South Asian woman, thick black wavy hair past shoulders, dark brown eyes, worn brown leather jacket over cream shirt) stands frozen on a rain-soaked Tokyo rooftop at 2am. She grips a crumpled letter in both hands. Medium shot from slight low angle. Rain backlit by neon pink and blue signage below. Desaturated teal-and-orange colour grade. Cinematic 35mm film look. Shallow depth of field, rooftop edge blurred behind her.

Lighting Keywords That Elevate Every Image

  • Golden hour — warm, soft, long shadows
  • Neon backlit — dramatic rim light, cyberpunk feel
  • Single candle — intimate, warm, sharp shadows
  • Overcast diffuse — flat, melancholic, soft
  • Volumetric fog — mysterious, depth layers
  • Hard midday sun — harsh shadows, bleached feel
  • Underlit from below — sinister, villain energy
  • Silhouette against fire — epic, dramatic

04 Character DNA — The Key to Visual Consistency

The biggest problem with AI storyboards is character inconsistency — your hero looks different in every scene. GoroSolution's Continuity Lock™ helps, but your Image Prompts must include a Character DNA block in every scene featuring that character.

Build Your Character DNA Once, Copy Everywhere

Write a single-paragraph description that you paste into every scene with that character. Be extremely specific:

Character DNA example — "Kai"
Kai (34yo East Asian man, sharp angular jaw, short undercut black hair with silver threads at temples, narrow dark brown eyes with deep-set intensity, lean athletic build, 6ft, olive-toned skin with a jagged scar from left eyebrow to cheekbone, always wearing his signature long black wool coat, dark grey turtleneck, tactical boots)

Elements of a Strong Character DNA

  • Age and ethnicity — be precise, not vague
  • Hair — colour, length, texture, style
  • Eyes — colour, shape, expression quality
  • Skin tone — specific description, not just a label
  • Build and height
  • Distinguishing marks — scars, tattoos, birthmarks
  • Signature outfit — what they wear in most scenes
Save your Character DNA in a text file. Paste it at the start of every image prompt where that character appears. Consistency across 10 scenes becomes effortless.

05 Writing the Video Prompt (Veo3)

The Video Prompt instructs AI video engines on how to animate your scene as an 8-second clip. GoroSolution uses this for Veo3 and similar engines. Structure it as a timed shot breakdown:

The 3-Beat Video Structure

Strong video prompt
[0–2s] OPEN: Kai stands at the bridge railing, back to camera, looking at the burning city below. Static wide shot. Wind moves his coat. [2–5s] RISE: He slowly turns. Medium shot, face coming into frame — eyes reflect the fire. Camera slowly pushes in. [5–8s] PEAK: Close-up on his hands — he drops a photograph into the flames below. Hold as it burns. Slow pull-back into black.

Video Prompt Rules

  • Always use timestamps: [0–2s], [2–5s], [5–8s]
  • Label the emotional arc: OPEN → RISE → PEAK
  • Include camera movement (push in, pull back, pan left, tilt up, static)
  • Describe character motion precisely — not "he moves" but "he slowly turns to face camera, jaw tightening"
  • Include audio cues — rain, silence, music swell, distant explosion
  • End with what the final frame looks like — this feeds directly into the Last Frame field

06 The Last Frame Field

The Last Frame is a visual handoff — it describes the final 1–2 seconds of your scene, and becomes the opening frame of the next scene's Image Prompt. This is what creates cinematic continuity across your storyboard.

Last Frame example
Kai standing at the bridge railing, side profile, long black coat soaked with rain, hands gripping the railing, looking off-screen left toward the burning skyline. Warm orange fire glow from left, cold blue night sky from right. Medium shot. His expression is unreadable.

The next scene's Image Prompt should open with: "Picking up from Kai at the bridge railing in the rain, he now turns and begins walking toward the camera..."

Always fill in Last Frame when using Continue Story. It tells the AI exactly how to start the next generated scene, preventing visual jumps.

07 Visual Style & Scene Mood

These two selectors in the left sidebar are powerful modifiers. The AI applies them across all generated scenes. Choose carefully — they define the visual DNA of your entire storyboard.

Visual Styles — What They Do

  • Cinematic — Widescreen, film grain, cinematic colour grading, depth of field. Best for drama, thriller, action.
  • Cyberpunk — Neon-lit, rain-soaked, dystopian urban. Blue/purple/pink palette. High contrast.
  • Anime — Japanese animation style, expressive characters, cel shading, dramatic composition.
  • Oil Painting — Classical painting texture, rich brushstrokes, warm palette. Great for historical or fantasy.
  • Dark Fantasy — Moody, gothic, deep shadows, medieval/magical elements.
  • Photorealistic — Closest to real photography. Detailed textures, natural lighting. Best for grounded stories.
  • Comic Book — Bold outlines, flat colours, dynamic angles, speech-bubble-ready compositions.
  • Watercolor — Soft edges, delicate colour bleeds, illustrative quality. Great for romance or indie storytelling.

Scene Moods — What They Do

  • Dramatic — High contrast, tense compositions, charged atmosphere
  • Dark — Low key lighting, shadow-heavy, oppressive atmosphere
  • Epic — Grand scale, hero framing, sweeping compositions
  • Romantic — Soft light, warm tones, intimate framing
  • Horror — Unsettling angles, creeping shadows, unnatural stillness
  • Serene — Open spaces, soft light, calm compositions
Mix and match intentionally. "Anime" + "Horror" creates something very different from "Cinematic" + "Horror". Experiment with combinations that match your story's unique tone.

08 Writing for Continuity

Continuity is the art of making scenes feel connected — like they exist in the same world, with the same characters, at the same moment in time. GoroSolution's Continuity Lock™ helps, but your prompts are the foundation.

The Continuity Note Field

Every scene in the Edit modal has a Continuity Note field. Use it to describe the visual link between this scene and the one before:

  • "Picks up directly from the rooftop — same rain, same neon glow, Kai still in wet coat"
  • "Morning after — Kai now wears a fresh white shirt, sun through window, relaxed"
  • "Cut to exterior — same building seen from street level, still night"

Continuity Checklist for Every Scene

  • Does the character look the same? (Use identical Character DNA)
  • Does the time of day make sense? (Day → Night → Day transitions)
  • Does the outfit match? (Unless there's a reason to change)
  • Does the environment connect? (Same location = same background details)
  • Does the emotional state carry over? (Or is there a clear reason for change)

09 Img2Img Prompts (Pro & Basic)

Image-to-Image (Img2Img) lets you upload a reference image and have the AI generate a new scene based on that visual. Available on Pro and Basic plans, Ultra HD phase only.

What to Upload as Reference

  • A previous generated scene — to maintain location or character look
  • A concept sketch — AI will refine and render it cinematically
  • A real photo — AI adapts the composition into your chosen style
  • A mood board image — to force a specific lighting or colour grade

Influence Slider (10%–100%)

  • 80–100% — Stays very close to your reference. Good for location consistency.
  • 50–70% — Balances reference and prompt. Most flexible setting.
  • 20–40% — Uses reference loosely. More creative interpretation. Good for style transfer.

Best Practice

When using Img2Img, write your Image Prompt as if describing what should change from the reference, not what should stay the same. The reference handles the stable elements automatically.

10 Real Examples — Good vs Bad

Genre: Sci-Fi Action

Concept — Too vague
Two soldiers fight on a spaceship.
Concept — Specific and visual
Commander Reyes (40yo Latina woman, silver-streaked black hair in a tight braid, cybernetic left arm glowing blue, scarred face, wearing a torn white combat suit) fights hand-to-hand against an AI soldier aboard a crumbling deep-space freighter. Emergency red lighting, decompression alarms, sparks. The ship is tilting 30 degrees. She's losing but won't stop. Filmed like Aliens meets Gravity — practical, brutal, claustrophobic.

Genre: Period Drama

Image Prompt — Too brief
A queen sits on a throne looking powerful.
Image Prompt — Cinematic
Queen Isadora (52yo Black woman, natural grey-streaked locs piled high with gold thread, deep brown skin, high regal cheekbones, wearing a blood-red gown with floor-length gold embroidery) sits rigid on an obsidian throne, hands gripping armrests, eyes fixed forward with cold fury. Wide shot from low angle — throne elevated, massive stone hall behind her. Single shaft of morning sunlight cuts through stained glass from above left, painting her in fractured gold and red. Shadow falls over half her face. Painterly oil-style rendering. 17th century European aesthetic.

11 Common Mistakes to Avoid

  • Describing emotion instead of visuals. "She feels sad" doesn't generate a frame. Write "her eyes are red-rimmed, jaw clenched, shoulders turned inward" instead.
  • Skipping the Character DNA. Without a DNA block, your character changes face every scene.
  • Not specifying lighting. Lighting is 50% of any image. Always include it.
  • Using abstract genre labels only. "Cyberpunk style" alone is weak. Describe the specific visual elements you want.
  • Leaving Story Beat empty. The story beat drives scene generation. "The hero makes a decision" is better than nothing.
  • Generating all images before reviewing scenes. First generate text scenes, review and edit them, then generate images. Saves credits.
  • Not using Last Frame. Without it, Continue Story has no visual anchor for the next scene.
  • Using "a man" or "a woman" without description. AI defaults to generic. Always describe your character specifically.

12 Pro Tips for Power Users

  • Build a prompt library. Save your best Character DNAs, lighting descriptions, and setting blocks. Reuse them across projects.
  • Use Act structure intentionally. Act 1 scenes: lighter colour grade, more open compositions. Act 2: tighter framing, higher contrast. Act 3: maximum tension or release.
  • Reference real films. "Lit like There Will Be Blood" or "composition like Parasite's staircase scene" — the AI understands cinematic references.
  • Use Custom Scene for pivotal moments. AI-generated scenes are great for volume, but hand-write Image Prompts for your key emotional moments.
  • Regenerate selectively. If one scene's image is wrong, edit just that Image Prompt and regenerate that single scene. Don't redo the whole board.
  • Set Aspect Ratio first. 16:9 for film/TV, 1:1 for social media, 9:16 for mobile/TikTok. This changes how scenes are composed by the AI.
  • Combine styles in your image prompt. "Cinematic photography with anime-style colour saturation and comic-book hard shadows" creates something genuinely unique.
  • Use the JSON export. Your full storyboard in JSON includes all prompts. Use it as a master prompt library for video generation tools like Veo3 or Kling.

Ready to generate your first storyboard?

Now that you know how to write powerful prompts, open the app and put it into practice.

Open Storyboard AI