What is your goal?
rss to instagram real using my own voice from eleven labs
What is the problem & what have you tried?
Title: JSON2Video “Movie duration cannot be zero” error when using Create a Movie from Template ID (Make.com)
Hi everyone,
I’m trying to build an automation in Make that creates Instagram reels automatically. The intended workflow is:
RSS feed → AI generates captions and script → ElevenLabs generates voice narration → JSON2Video creates the video → Instagram posts the reel.
Everything works until the JSON2Video step.
I’m using the Create a Movie from Template ID module, but every time the scenario runs I get this error:
RuntimeError: Movie duration cannot be zero
From the logs it looks like JSON2Video accepts the project but fails during rendering.
What I’ve already tried:
• Verified all API connections (OpenAI, ElevenLabs, JSON2Video, Instagram)
• Compared my scenario with the original “Automated Social Media Video Machine” blueprint
• Checked that my Parse JSON structure matches the AI output
• Tested both scene-based JSON structures and caption/voice structures
• Built a simplified template with only one scene, a fixed image URL, fixed text, and an explicit duration
Example template I tested:
{
"resolution": "instagram-story",
"scenes": [
{
"duration": 8,
"elements": [
{
"type": "image",
"src": "https://images.unsplash.com/photo-1500530855697-b586d89ba3ee",
"duration": 8
},
{
"type": "text",
"text": "Test video",
"style": "003",
"duration": 8
}
]
}
]
}
Even with this simple template I still get the same error.
My goal is simply to automate short reels like this:
RSS article → AI script → ElevenLabs narration → JSON2Video reel → Instagram
Has anyone successfully used JSON2Video Create a Movie from Template ID in Make and encountered this error? If so, what usually causes the “movie duration cannot be zero” problem?
Any guidance would be appreciated.
Error messages or input/output bundles
RuntimeError: Movie duration cannot be zero
From the logs it looks like JSON2Video accepts the project but fails during rendering.
What I’ve already tried:
• Verified all API connections (OpenAI, ElevenLabs, JSON2Video, Instagram)
• Compared my scenario with the original “Automated Social Media Video Machine” blueprint
• Checked that my Parse JSON structure matches the AI output
• Tested both scene-based JSON structures and caption/voice structures
• Built a simplified template with only one scene, a fixed image URL, fixed text, and an explicit duration




