Video Content for Social Media

Generate short-form video clips from text prompts — aspect ratios, duration, visual styles, and prompting strategies for social platforms.

Which video model should you use?

Platform goal	Recommended model	Why
TikTok / Reels (vertical)	`truefusion-video-pro`	Native 9:16, cinematic quality
YouTube Shorts	`truefusion-video-pro`	Flexible aspect ratios
LinkedIn / Twitter	`mera`	Professional, cinematic look
Stylized / artistic	`lumo`	Unique visual style
Google-quality	`veo-3.1`	Photorealistic motion

Aspect ratios by platform

Platform	Aspect ratio	Input value
TikTok, Instagram Reels	9:16 (vertical)	`"9:16"`
YouTube Shorts	9:16 or 1:1	`"9:16"`
YouTube landscape	16:9	`"16:9"`
Instagram feed	1:1	`"1:1"`
Twitter / X	16:9	`"16:9"`

curl -X POST https://api.skytells.ai/v1/predictions \
  -H "x-api-key: $SKYTELLS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "truefusion-video-pro",
    "input": {
      "prompt": "A barista pours latte art in a sunlit café, slow motion, warm tones, cinematic",
      "duration_seconds": 5,
      "aspect_ratio": "9:16",
      "fps": 24
    }
  }'

Typical response time: 30–90 seconds depending on model and duration.

Good social video prompts have four elements:

Subject + Action + Setting + Style

[Subject] [doing what] in [setting], [style], [mood]

High-performing templates

Product showcase:

A [product] on a [surface], rotating slowly, studio lighting, 
product photography style, clean white background

Lifestyle:

A person [activity] in [location], golden hour, handheld camera,
lifestyle brand feel, warm color grade

Nature / ambient:

[Scene description], timelapse, 4K, cinematic, [mood] atmosphere

Abstract / motion:

Abstract [shapes/colors] flowing and morphing, [aesthetic] style,
seamless loop, [color palette]

What makes a strong prompt

Element	Weak	Strong
Subject	"a person"	"a chef in a white apron"
Motion	"moving"	"slow motion pour"
Style	none	"cinematic, shallow depth of field"
Lighting	none	"golden hour, warm backlight"
Color	none	"muted earth tones, film grain"

Batch generation — generate 5 variations

Use cURL in a simple shell loop to generate multiple variations quickly:

#!/bin/bash
PROMPTS=(
  "Ocean waves crashing at sunset, slow motion, 9:16, cinematic"
  "City streets at night, rain, bokeh lights, vertical"
  "Coffee steam rising in morning light, macro, warm tones"
  "Forest path in autumn, leaves falling, handheld, peaceful"
  "Product unboxing on marble surface, satisfying, clean"
)

for prompt in "${PROMPTS[@]}"; do
  curl -s -X POST https://api.skytells.ai/v1/predictions \
    -H "x-api-key: $SKYTELLS_API_KEY" \
    -H "Content-Type: application/json" \
    -d "{\"model\": \"truefusion-video-pro\", \"input\": {\"prompt\": \"$prompt\", \"duration_seconds\": 5, \"aspect_ratio\": \"9:16\"}}" \
    | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['id'], d['status'])"
done

Polling for completion

Video predictions take time. Poll until status === "succeeded":

ID="pred_vid_abc123"
while true; do
  STATUS=$(curl -s https://api.skytells.ai/v1/predictions/$ID \
    -H "x-api-key: $SKYTELLS_API_KEY" \
    | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['status'])")
  echo "Status: $STATUS"
  [ "$STATUS" = "succeeded" ] || [ "$STATUS" = "failed" ] && break
  sleep 10
done

Downloading your video

Output URLs expire in 24 hours. Download immediately:

URL=$(curl -s https://api.skytells.ai/v1/predictions/$ID \
  -H "x-api-key: $SKYTELLS_API_KEY" \
  | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['output'][0])")

curl -L "$URL" -o "social_video.mp4"
echo "Saved to social_video.mp4"

Workflow recommendation

Generate in bulk during off-hours. Review and select the best clips. Add audio (next module). Post during peak engagement times.

Summary

Match aspect ratio to platform — 9:16 for TikTok/Reels, 16:9 for YouTube
Use the prompt formula: Subject + Action + Setting + Style
Generate multiple variations — AI results vary, pick the best
Download outputs before 24-hour expiry
Batch generate with shell scripts or automations

NextAI Voiceovers & Music

Video Content for Social Media

On this page