Fantasy Prompt Lab
Director's Cut
A unified tool for directing advanced AI video and images, and generating consistent character concepts.
Character Visuals
Who is your character? Think about their species, age, build, and what makes them visually unique.
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Scene & Composition
Where is your character and what are they doing? Define the overall look and camera work for the scene.
Setting & Action
Establish the context. What is the environment, who is the focus, and what is the primary activity?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Artistic Style
Define the overall visual aesthetic. Is it a painting, a render, or something else?
Would you like to add your own?
Cinematography
How should the scene be filmed? Control the camera's shot type, angle, and movement.
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
Would you like to add your own?
🎬 Narrative & Performance
How does the story unfold in this shot? Detail the sequence of actions and the character's emotional journey.
Action Sequencer
For video, this creates a step-by-step sequence and overrides the general 'Action' selection.
Character Performance Arc
Define the character's emotional change from the beginning of the shot to the end.
Gaze & Focus
Direct the character's attention to avoid the default camera stare.
🎤 Integrated Sound Design
What does this scene sound like? Layer in dialogue, background noise, and music to build a complete world.
Dialogue
Would you like to add your own?
Layered Sound Design
Add and describe individual audio layers like music, ambience, and specific SFX.
Would you like to add your own?
Would you like to add your own?
🚫 Refinements
What should the AI specifically avoid? Use negative prompts to fix common errors and steer the generation away from unwanted elements.
Quality Fixes: Adds terms to prevent blurriness, artifacts, and low resolution.
Anatomy Fixes: Adds terms to prevent common issues like extra limbs and distorted features.
Prompt Output & Controls
Technical Parameters
Core Concepts for Consistent Characters
Understanding these foundational principles is key to success in AI character generation.
Multimodal Input
Advanced AI models often use both a detailed text prompt and a strong visual reference image. Create a high-quality reference image of your character first to anchor consistency.
Character Consistency
For a consistent character across multiple outputs (images or video clips), use the same reference image AND the same detailed text description in every prompt. For sequential video shots, leverage platform-specific continuity features.
The SDAS Prompting Model
This tool is built around an effective prompting strategy called the Subject-Detail-Action-Style (SDAS) model. By structuring prompts this way, we minimize ambiguity and give the AI clear, layered instructions, which is key to generating consistent and high-quality results.
- 1. Subject: First, we define the core subject (a half-orc male) and the overall medium (Unreal Engine render). This is the foundation.
- 2. Detail: Next, we layer on all the physical characteristics, from broad (muscular build) to specific (tattered clothing, facial tattoos). This paints a detailed picture for the AI.
- 3. Action: With the character defined, we now describe what they are doing (battling a goblin, shouting "For glory!"). This provides narrative context.
- 4. Style: Finally, we add cinematic and artistic instructions (dynamic action pose, cinematic lighting). This tells the AI how to frame and present the scene.