Find where the failure begins
Watch the clip once at normal speed, then inspect the problem section slowly. Check whether the source image already contains ambiguous edges or overlapping objects. A camera move that reveals an unseen side may cause a different failure from a stationary object whose surface changes colour.
For I2V, compare against the actual source. If the input shows only half of an object, the model must infer the rest when it moves. Preparing a clearer source can be more effective than rewriting the entire prompt.
Run a focused retry
- Keep the same source and simplify the requested motion.
- Remove competing camera instructions and large rotations.
- Generate another short attempt and compare the same troublesome detail.
- If the failure remains, change the source composition before adding more prompt detail.
Example repair prompt
Preserve the source object's outline, colour and visible details. The object stays still. Use only a small straight camera push-in, with steady lighting and no scene transition. Keep the full object inside the frame throughout the shot.
Know when to stop retrying
If a strict product detail still changes, use a still-image animation or real footage for that section. A shorter clean interval may be usable, but do not hide a misleading product transformation behind a fast edit. Keep the failed examples and their settings if you contact support; a precise description such as “the handle splits at the midpoint” is more actionable than “bad quality.”
REVIEWED JOLLYAI IMPLEMENTATION
The workflow behind the guide
One uploaded image + motion brief → image and text conditioning → H3 FL2VA turbo sampling → video/audio decoding → combined video file.
- Inputs, presets & access
- One image; aspect follows the source. Limited free access and longer PRO presets. Current ordinary pixel budget: 0.4 MP.
- Implementation note
- Local H3 turbo workflow. Six sampling steps; simple scheduler; turbo adapter; video/audio shifts 12/3. Current environment: pruned W4A8 FL2VA weights with a 4B encoder and matching projection. R2V uses a separate reference checkpoint. These settings describe the reviewed JollyAI deployment, not every H3 implementation.
Configuration reviewed September 20, 2026. Your current playground controls and plans page determine availability. Prompt recipes above are suggestions, not guarantees or claimed test results.
Open Image to video ↗Sources and useful next steps
The implementation notes describe JollyAI's reviewed local integration. Upstream capabilities and licences may differ from the settings available here.
- MiniMax H3 upstream project and model information ↗
- MiniMax H3 tool overview on JollyAI
- Earlier H3 ComfyUI workflow snapshot and downloadable JSON — check the newer configuration notes above when reproducing it.
- JollyAI content policy and service terms
Found an outdated setting or a reproducible problem? Contact JollyAI with the workflow and job reference. Keep account tokens and private input files out of public reports.