Input Types
Generate content from text, images, or video.
PowerPost supports three input types for content generation. Each produces platform-optimized captions with hashtags.
Text
The simplest option — describe what you want to post about.
{
"prompt": "We just launched our new mobile app with AI features",
"post_types": ["instagram-reel", "tiktok-video"],
"research_mode": "regular"
}The prompt field is required (at least 3 characters, up to 2,000 tokens). PowerPost researches the topic and generates tailored content for each platform.
Images
Upload images via the Upload Media endpoint, then pass the returned media_ids here. PowerPost analyzes the images and generates captions. Optionally provide text for additional context.
Supported formats: JPEG, PNG, WebP
{
"media_ids": ["a1b2c3d4-...", "e5f6g7h8-..."],
"prompt": "Behind the scenes at our product launch event",
"post_types": ["instagram-feed", "facebook-post"],
"research_mode": "regular"
}media_ids— Array of uploaded media IDs (up to 10 images)prompt— Optional context to guide the generated content
Use the optional prompt field to steer the AI — for example, highlighting a specific detail in
the image or providing context that isn't visually obvious.
Video
Upload a video via the Upload Media endpoint, then pass the returned media_id here. PowerPost analyzes the video using AI video understanding and generates captions based on the content.
Supported formats: MP4, MOV
{
"media_ids": ["f9g0h1i2-..."],
"prompt": "Focus on the key insights about productivity",
"post_types": ["tiktok-video", "youtube-short", "x-post"],
"research_mode": "deep"
}media_ids— Array with the uploaded video media IDprompt— Optional context to focus the generated content
Video input requires additional credits for video analysis. Credit cost also scales with video duration.
Comparison
| Input Type | Best For | Processing |
|---|---|---|
| Text | Quick ideas, announcements | Immediate |
| Images | Product photos, event shots, designs | Image analysis |
| Video | Podcasts, presentations, clips | AI video analysis |
Post types
Pass post_types to choose platforms and surfaces. Full list and media/caption limits: Post types.
Multiple types on the same platform produce one shared caption for that platform.
Output format
Outputs are keyed by platform — one caption per unique platform derived from your post types:
{
"tiktok": "🌙 Dark mode activated! POV: your eyes at 2am... #darkmode #tech",
"instagram": "✨ Dark mode is here!\n\nYour late-night scrolling just got easier... #DarkMode #AppUpdate",
"youtube": {
"title": "We Just Shipped Dark Mode",
"description": "Dark mode is finally here across all our apps... #darkmode #tech"
}
}YouTube outputs have a separate title and description since those are distinct fields on the
platform. All other platforms output a single string.