Movie Creator · skill pública de Open Skill

Movie Creator

AI Video Generator (Custom Duration & Style Control) Transform any text prompt into high-quality, realistic AI video clips with full control over length and video parameters. Key Features: Custom Duration: Automatically detects requested seconds (e.g., 10, 20s, 30s) and generates exact length clips. Text-to-Video Engine: Creates detailed visual scenes, cinematic animations, and action sequences from simple text descriptions. Multiple Aspect Ratios: Supports Landscape (16:9), Portrait (9:16), and Square (1:1) formats for social media and general use. Iterative Editing: Easily re-generate, refine visual style, or adjust scene details based on user feedback.

@Bhatti

Custom

63 45 55

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4241498/2026-10-08/anbg8y2b.png

# AI Video Generator

## When to use
- The user wants to turn a text prompt into a video clip (no input image provided).
- The user asks for a video of a specific length, e.g. "generate a 10-second video of a city at night", "make a 5-second clip of waves crashing".
- The user wants to redo or adjust the last generated video.

## Steps
1. Read the user's text prompt and **extract the desired duration in seconds** if they mention one (e.g. "10 seconds", "5s", "half a minute"). If the user gives a duration in another unit (minutes, etc.), convert it to seconds. If no duration is given, use a sensible default (e.g. 5 seconds).
2. Call `generate_video` with:
   - `model='t2v'` (text-to-video; pass it explicitly because there is no input image — omitting `model` with no input image is rejected).
   - `prompt`: the user's text prompt describing the scene, motion, and style.
   - `duration`: the number of seconds extracted in step 1 (this is the key requirement — always honor the user's requested length).
   - `orientation`: `landscape` (16:9) by default, or `portrait` (9:16) / `square` (1:1) if the user asks for a specific format.
   - `resolution`: `720P` by default (auto-clamped by the model if exceeded).
   - `name`: a short display title for the clip (e.g. "City at night — 10s").
   - `entity_type`: `final_videos`.
3. On success, summarize the result in natural language (what the clip shows, its length, orientation). Do not paste large blocks of URLs.
4. If the user wants changes (different length, style, scene), regenerate with the updated prompt/duration.

## Notes
- Video generation does not support NSFW/adult content on any channel — politely decline such requests.
- If the requested duration exceeds the model's limits, the platform auto-clamps it; mention the final length to the user.
- Report failures honestly; never fabricate a video URL.