How to Use Seedream and Seedance in FLORA for Cinematic Workflows

Introducing ByteDance’s Latest Models in FLORA

ByteDance—the company behind TikTok and CapCut—recently introduced two powerful models now available inside FLORA:

• Seedream 3.0: a text-to-image model that generates high-resolution images in under 3 seconds, with strong bilingual text handling and the ability to create complex, cinematic scenes.

• Seedance 1.0: a text-to-video and image-to-video model designed for motion stability, cinematic quality, and multi-shot storytelling.

Used together, these models unlock a faster way to go from concept to moving sequence—all inside FLORA.

Mastering Prompts with Seedream 3.0

As a text-to-image model, Seedream depends on clear, structured prompts. From ByteDance’s own documentation, here’s a bite-size guideline for building effective inputs:

1. Start with style + medium (e.g. “experimental film still”)

2. Add subject description in natural language

3. Include environmental context (location, setting, atmosphere)

4. Refine with technical photography cues (camera model, lens, lighting)

5. Finish with mood or emotional tone

Example 1: Cinematic Realism

Prompt: “Experimental film still, person brushing teeth, dimly lit bathroom, Sony Venice 2 with Canon K35 lens, medium wide shot, contemplative mood.”

Result: A filmic still that captures cinematic lighting, composition, and tone in seconds.

Example 2: Black-and-White Drama

Prompt: “Hyperrealistic black and white digital photograph, two human shadows, Leica M6 with 35mm lens, ILFORD 400 film, shot in a narrow alley.”

Result: A dramatic, high-contrast image with analog-style authenticity.

Cinematic Motion with Seedance 1.0

Seedance builds on Seedream by animating stills or creating videos directly from text. Its strength lies in camera awareness and motion fidelity.

Example 1: Tracking Shot

Prompt: “Tracking medium shot of a pigeon weaving through a busy crowd.”

• Smooth subject motion

• Realistic environmental rendering

Example 2: Environmental Variation

Using the same prompt but swapping in “subway station” for the environment produces a new scene while keeping shot type, pacing, and subject motion consistent.

Example 3: Camera Angles

Seedance understands directives like “low angle shot” or “over-the-shoulder close-up.” It preserves cinematic framing and subject movement across shots.

Multi-Shot Storytelling in a Single Prompt

One of Seedance’s standout features is native multi-shot generation. You can structure a sequence within brackets:

Prompt:
“[Medium shot: a man walking through a market] → cut to [Over-the-shoulder shot: same man entering a café] → cut to [Extreme close-up: his eyes glancing upward].”

• Works best with up to three shots per prompt

• Allows seamless transitions between perspectives

• Unlocks narrative storytelling from a single generation

Why It Matters for Creative Professionals

With Seedream and Seedance inside FLORA, creative teams can:

• Prototype editorial and cinematic visuals faster than ever

• Maintain style consistency across stills and video

• Experiment with professional camera logic directly in prompts

• Move from concept art to full motion sequences in one workflow

Try It Yourself in FLORA

Seedream + Seedance give you precision, speed, and cinematic control—all from text or image prompts. Whether you’re building ad campaigns, film storyboards, or experiential visuals, these models let you explore at production speed.

👉 Sign up for FLORA and start creating with Seedream + Seedance