Automated UGC Review Video Builder

UGC Ad Video Automation (n8n + Telegram + AI Video SDK)
Goal

We want to build a stable and sustainable automation system that generates realistic UGC-style ad videos for Shopify products.

Important:
At this stage, we do NOT want Shopify publishing, ad uploading, or posting automation.
We only want high-quality ad assets (video + thumbnail + caption).

1- Final Outputs (per request)

For each request, the system must generate:

1 UGC-style ad video

Format: MP4

Ratio: 9:16 (vertical)

Duration: 15 seconds

1 thumbnail / cover image

Short ad copy + CTA text

(Optional) Prompt / seed log for reproducibility

2- Video Style Requirements (VERY IMPORTANT)

The videos must look like real user-generated content, not AI or studio-made.

Style: Ultra-realistic UGC

Camera: Smartphone recording, slight handheld shake

Lighting: Natural home daylight

Environment: Real home / desk / office-like setting

Human: Looks like a real person (no obvious AI artifacts)

No cinematic look, no studio lighting, no watermarks

Typical 15s structure:

0–3s: Problem (discomfort, pain, frustration)

3–10s: Product usage

10–15s: Relief + CTA

CTA example:
“Relax and keep going”

3-How the System Will Be Used

A user sends ONE message to a Telegram bot, for example:

Product: Neck massager
Target audience: Desk workers
Style: Realistic UGC
Video length: 15 seconds
Lighting: Natural home light
Camera: Phone, handheld
CTA: Relax and keep going


(Optional) The user may also send 1 reference image via Telegram.

The automation generates the video and sends the final MP4 back via Telegram.

4- Technical Expectations

Automation platform: n8n (cloud or self-hosted)

Trigger: Telegram bot (message + optional photo)

Prompt generation: LLM (OpenAI / Gemini or equivalent)

Image generation: SDK-based (NanoBanana / Banana / equivalent)

Video generation: SDK-based (VEO / Runway / Kling or equivalent)

Delivery: Telegram sendVideo or sendDocument

Output files should be stored temporarily or permanently (S3 / R2 / Drive optional)

5- Critical Technical Issue to Solve

We currently have a known issue that must be solved properly:

Telegram image handling fails with errors like:

“File URL not found”

YOUR_BOT_TOKEN placeholder not replaced

Incorrect file_id → getFile → file_path mapping

Expected solution:

Correct handling of Telegram photo events

Proper getFile call using the real bot token

Correct construction of:

https://api.telegram.org/file/bot<REAL_BOT_TOKEN>/<file_path>


Or passing the image as binary directly to AI nodes (preferred if possible)

The system must work reliably, not just once.

6- Prompt Standard (Automation-Friendly)

The system should internally convert each request into a structured format like:

{
"product": "",
"target_audience": "",
"style": "realistic UGC",
"duration_sec": 15,
"lighting": "natural home light",
"camera": "handheld phone",
"scene": "",
"hook": "",
"cta": ""
}


If required data is missing, the system may ask maximum one clarification question.

7- What We Expect From the Freelancer

A working n8n workflow (importable JSON)

Clear setup instructions (credentials, API keys, SDKs)

One live demo:

Telegram message → video generated → video sent back

Basic logging / retry logic for failures

8- What We Do NOT Need (for now)

❌ Shopify product page integration
❌ Ad account publishing
❌ Scheduling / posting to social platforms
❌ Over-engineered dashboards

Focus only on clean, realistic ad video generation.

9- Budget & Mindset

We are looking for someone experienced with:

n8n

Telegram bots

AI image/video SDKs

We prefer a clean, maintainable solution over quick hacks.

If you are experienced with n8n + Telegram + AI video pipelines, please apply with:

Relevant past work

Which video SDK you recommend and why

Estimated delivery time