muapi-ugc-video-factory

Turn a person photo + a product photo + an optional script into a vertical 9:16 UGC-style video ad.

Paste into Claude, ChatGPT or Cursor.

Read the source134 lines
muapi-ugc-video-factory/SKILL.md134 lines6.5 KBpushed 100d agoRawView on GitHub
1---
2slug: muapi-ugc-video-factory
3name: muapi-ugc-video-factory
4version: "1.0.0"
5description: Turn a person photo + a product photo + an optional script into a vertical 9:16 UGC-style video ad. Generates a lifestyle hero image (Nano-Banana Pro Edit), then animates it with native audio using Seedance 2.0 VIP image-to-video.
6acceptLicenseTerms: true
7---
8 
9 
10# UGC Video Factory
11 
12**Turn a person photo + product photo (+ optional script & environment) into a vertical 9:16 UGC-style video ad with native dialogue audio.**
13 
14A three-stage pipeline:
151. **GPT** writes a director-grade ultra-realistic lifestyle photography prompt from your inputs.
162. **Nano-Banana Pro Edit** fuses the person + product into a single hero photo (1K, 9:16).
173. **Seedance 2.0 VIP Image-to-Video** animates the hero photo into a 10s vertical UGC clip with synced spoken audio.
18 
19## Inputs
20 
21| Name | Type | Required | Default | Description |
22|:---|:---|:---|:---|:---|
23| `person` | image_url | yes | — | Photo of the person who will appear in the ad (face + upper body works best). |
24| `product` | image_url | yes | — | Clear photo of the product (preferably on neutral background, logo/text legible). |
25| `script` | text | no | `Okay… first of all, ship happens. And this hat is honestly my favorite. It also comes in navy and black, so you can pick your vibe.` | The exact line the on-screen person will say (kept short — 1–2 sentences fit 10s comfortably). |
26| `environment` | text | no | `study room, laptop in front of it` | Scene / context where the person is using the product (e.g. "bathroom mirror, morning routine", "coffee shop window seat"). |
27 
28If `person` or `product` is missing, ask the user to upload them (`muapi upload file <path>`) or offer to generate placeholders before continuing.
29 
30 
31## Steps
32 
33Run the three steps sequentially — each step's output feeds the next.
34 
35### Step 1 — Director Prompt (GPT)
36 
37Use a GPT model (`gpt-5.1` or whichever chat model is available to the executing agent) with **temperature 0** and **max ~200 tokens** to produce the hero-image prompt.
38 
39**System prompt:** `You are a helpful assistant.`
40 
41**User prompt** (substitute `{{person}}`, `{{product}}`, `{{environment}}`):
42 
43```
44Uploaded images are being analyzed. Ultra-realistic lifestyle photography with {{person}} and {{product}} and {{environment}}.
45 
46If the product is wearable (e.g., hat, glasses, hooded sweatshirt), the person wears the product naturally.
47 
48If the product is carried in the hand (e.g., cream, bottle, thermos), the person holds the product naturally.
49 
50The product is clearly visible and is the main focus of the image. The logo or text on the product must be legible.
51 
52The person has a natural and modern look with a minimalist style.
53 
54The scene is consistent with the context of the product's use: {{environment}}.
55 
56Lighting: soft natural daylight.
57Background: clean, aesthetic, slightly blurred (shallow depth of field).
58Style: high-end commercial lifestyle photography, realistic textures, 4K quality, vertical 9:16 composition, social-media advertising style. The background and environment should be appropriate to the product (e.g. a woman with a serum could be at home). The person's facial details and the product must remain unchanged.
59```
60 
61Capture the GPT response as `{{step1_prompt}}`.
62 
63### Step 2 — Hero Image (Nano-Banana Pro Edit)
64 
65Submit a `muapi image edit` call against the `nano-banana-pro-edit` model:
66 
67- **Reference images** (`image_urls`): `[ {{person}}, {{product}} ]` — order matters; person first.
68- **Prompt**: `{{step1_prompt}}` from Step 1.
69- **Aspect ratio**: `9:16`
70- **Num images**: `1`
71- **Resolution**: `1K`
72- **Output format**: `jpeg`
73 
74Capture the resulting image URL as `{{hero_image}}`. Briefly show it to the user for approval before kicking off the video step.
75 
76### Step 3 — UGC Video (Seedance 2.0 VIP Image-to-Video)
77 
78Submit a `muapi video from-image` call against **`seedance-2-vip-image-to-video`** (or the `-fast` variant if the executing agent wants lower latency).
79 
80- **Start image**: `{{hero_image}}` from Step 2.
81- **Aspect ratio**: `9:16`
82- **Duration**: `10` seconds.
83- **Generate audio**: `true` (native dialogue).
84- **CFG scale**: `0.5`
85- **Negative prompt**: `blur, distort, low quality`
86- **Prompt** (substitute `{{script}}`):
87 
88```
89Create a 10-second vertical UGC-style video (9:16).
90 
91A person is interacting naturally with their setting and product.
92 
93The product is used naturally:
94- If wearable → the person is wearing it.
95- If handheld → the person is holding or applying it.
96 
97The video is a single, uninterrupted shot. No cuts. No color changes. No text on screen.
98 
99The person looks directly at the camera with a relaxed and natural expression.
100They interact comfortably with the product using their hands (adjusting, holding, pointing).
101 
102They say in a natural, conversational tone:
103 
104"{{script}}"
105 
106Subtle hand gestures while speaking.
107End with a small smile or nod.
108 
109Style: authentic UGC, handheld phone feel, light natural movement, soft daylight, shallow depth of field, TikTok/Reels aesthetic.
110```
111 
112Poll the result with `muapi predict wait <request_id>` and download to the user's outputs directory.
113 
114## Notes
115 
116- VIP tier supports 9:16 and durations 4–15s; 10s is the sweet spot for a 1–2 sentence script.
117- Keep the script short — Seedance 2.0 will compress longer scripts and clip words.
118- Seedance VIP tolerates realistic human faces in references (unlike Chinese tier), making it the right choice for UGC.
119- If you want lower latency at the same quality, swap to `seedance-2-vip-image-to-video-fast`.
120- For multi-shot ads, generate several `{{hero_image}}` variations in Step 2 and animate each independently — Seedance VIP does not multi-image i2v at 9:16 + audio.
121 
122## Trigger Keywords
123 
124`ugc video factory`, `ugc video ad`, `person plus product video`, `talking product ad`, `ugc reel`, `lifestyle product video`, `vertical ugc video`
125 
126 
127---
128 
129## Notes for the Executing Agent
130 
131- This recipe is LLM-orchestrated: read each phase, gather any missing inputs from the user, then call `muapi` CLI commands. Run `muapi auth configure` first if `MUAPI_API_KEY` is unset.
132- For local files supplied by the user, upload them first: `muapi upload file <path> --output-json --jq '.url'`.
133- Substitute `{{input_name}}` placeholders with the user's actual inputs before issuing each call.
134- If the `muapi` CLI does not yet alias `nano-banana-pro-edit` or `seedance-2-vip-image-to-video`, fall back to the raw API: `curl -X POST https://api.muapi.ai/api/v1/<endpoint> -H "x-api-key: $MUAPI_API_KEY" -H 'content-type: application/json' -d '{...}'`, then poll with `muapi predict wait <request_id>`.

Discussion