Ugc Video

Orientation

Ask the user what format they want:

• 9:16 (vertical) — TikTok, Instagram Reels, YouTube Shorts (recommended for UGC)
• 16:9 (landscape) — YouTube, ads
• 1:1 (square) — Instagram feed

Default to 9:16 if they do not specify.

Brief

Ask the user about their product:

• Product name and what it is
• Key selling points (2-3 max)
• Target audience
• Tone: casual/fun, professional, excited, calm
• Any specific phrases or CTAs they want included

Phase 2: Script Generation

Write a UGC video script based on the brief. UGC scripts sound natural and unscripted.

Script Guidelines

Intro (hook) — 2-4 seconds:

• Start with a casual hook that grabs attention
• Examples: "Okay so I have to tell you about this...", "I just found the best [product category]...", "POV: you finally discover [benefit]..."
• Must feel spontaneous, not rehearsed

Product showcase — 4-8 seconds:

• Naturally introduce the product
• Show genuine enthusiasm, not salesy pitch
• Mention 1-2 key benefits conversationally
• Examples: "Look at this packaging though...", "And the best part is...", "I've been using it for a week and honestly..."

Present to User

Show the script to the user with clear labels:

INTRO: "[hook text]"
PRODUCT: "[product showcase text]"

Ask if they want to edit anything. Iterate until they approve.

Phase 3: Generate Missing Assets

If the user needs a character or product image generated:

Character Generation

Use xainflow_generate_image with:

• model: gemini-pro (highest quality for realistic people)
• aspect_ratio: match the chosen orientation (9:16, 16:9, or 1:1)
• prompt: Detailed description of a real person. Include:
  • Physical appearance (age, features, hair)
  • Clothing style (casual, trendy — matches UGC aesthetic)
  • Setting (home, kitchen, bathroom, outdoors — wherever product is used)
  • Expression (genuine smile, excited, natural)
  • Camera: medium close-up, looking at camera, natural lighting
  • IMPORTANT: Do NOT mention any product in the character prompt
• save_to_gallery: true

Show the generated character to the user. Ask if they want to regenerate or proceed.

Product Generation (if needed)

Use xainflow_generate_image with:

• model: gemini-pro
• aspect_ratio: 1:1 (product shots work best square)
• prompt: Clean product photo on simple background
• save_to_gallery: true

Phase 4: Build Workflow

Use xainflow_create_workflow with a complete flow_state containing all nodes and edges.

Workflow Name

Name it: "UGC - [Product Name]" (e.g., "UGC - Organic Ketchup")

Required Nodes

Build the workflow with these nodes. Positions assume 9:16 orientation — adjust group widths if landscape.

Use the user's chosen orientation for ratio (imageGenerator) and videoRatio (videoGenerator).

Use the user's chosen orientation to determine video duration:

• Short scripts (under 15 words): "4" seconds
• Medium scripts (15-30 words): "6" seconds
• Long scripts (30+ words): "8" seconds

GROUPS (7 groups):

All groups: zIndex: -1, fontSize: "lg"

NODES inside groups:

Edges

Character Description Prompt

Write a detailed visual description of the character for the prompt-char-desc node. This prompt is used to generate a high-quality reference image. Include:

• Physical appearance matching the character asset
• Setting/environment where the UGC video takes place
• Clothing and styling
• Natural pose and expression
• Lighting (natural, warm, casual)
• DO NOT mention any product

Video Prompt Guidelines

When setting prompts for video nodes:

• Include explicit motion (action verbs with direction/tempo)
• Include camera movement (slow pan, dolly, tracking shot, steadicam)
• With first frame (image-to-video): DON'T re-describe what's visible. Focus on what MOVES and HOW
• Motion verbs: billows, flutters, sways, cascades, drifts, sweeps, blooms, ripples
• For 4-5s: 1 action + 1 camera move
• For 6-10s: 2 beats (start → change)
• No meta-language ("a video of"), no filler ("beautiful", "stunning")
• Include speech in quotes for audio-enabled videos

Phase 5: Summary & Execution

Show Summary

Before executing, show the user:

UGC Video Workflow Summary:
- Character: [description or asset name]
- Product: [product name]
- Script:
  INTRO: "[intro text]"
  PRODUCT: "[product text]"
- Format: [9:16 / 16:9 / 1:1]
- Video model: Seedance 1.5 Pro (720p, with audio)
- Estimated credits: ~300-500

The workflow will generate:
1. Character reference image (150 cr)
2. Character intro image (150 cr)
3. Intro video clip (~48-96 cr)
4. Character + product image (150 cr)
5. Product video clip (~48-96 cr)

Ask for confirmation before executing.

Execute

Use xainflow_execute_workflow with the workflow_id.

Important: Video nodes are ASYNCHRONOUS. After execution:

1. Image nodes will complete immediately — report their results
2. Video nodes will return generation_id — tell the user videos are processing
3. Wait ~60 seconds, then poll each video with xainflow_get_video_status(generation_id)
4. Keep polling every 30 seconds until all videos complete or fail
5. Report final results with video URLs

Output

Save all generated assets to the output-folder ("UGC Videos"). Create the folder if it does not exist using xainflow_manage_folders.

Present final results:

UGC Video Complete!

Generated assets:
- Character image: [link]
- Intro video: [link] ([duration]s)
- Product composite image: [link]
- Product video: [link] ([duration]s)

Total credits used: [amount]