Assemble a finished clip from your own Prizmad assets: an ordered list of video/image clips (chat generations, Dashboard library, Products media or uploads) with a voice-over, background music (auto-ducked under the voice), captions and timed overlays.
Ordered timeline clips as a JSON array of {url, duration, start?, kind?} (1-12 items, each 0.5-600s, total <= 600s). "duration" is how long the clip plays; "start" trims into the source; "kind" is auto-detected (video or image). Clips play back-to-back in order. Use the URLs returned by the tools (chat generations, list_my_videos, upload_image) — the server verifies ownership.
Voice-over track from your assets (e.g. Text to Speech or your AI avatar audio). Music is ducked automatically under its timing.
Background music from your assets (e.g. Background Music). Loudness is lowered automatically while the voice-over is speaking.
Output frame. Vertical 9:16 is the default for Reels/Shorts/TikTok.
Timed caption words as a JSON array of {text, start, end} in seconds (max 5000). Use the timings from the transcription/text-to-speech step — do not retype the words.
Visual preset for the animated captions. Only used when captions are provided.
Timed overlays as a JSON array (max 8). Each item has "kind": "cta" | "price" | "review" | "text", optional "start"/"end" in seconds, plus "text" (cta/text), "title"/"price"/"showCta" (price) or "title"/"starCount" (review). Text is capped at 280 characters.
Default title used by price and review overlays when they do not set their own.
Default price used by the price overlay when it does not set its own, e.g. "$29.99".
Sign in to spend tokens
Clips (JSON) is required
Your generated video will appear here