Forusers without the dreamina CLI who are regular members. Only the opencli jimeng subcommand is allowed;forbidden dreamina image2video、frames2video、multiframe2video、multimodal2video and jimeng-cli-image2video。
The incremental motion prompt is provided by jimeng-prompt-image2video ;image-to-video / first-last frame / storyboard / all-in-one referenceis generated on the Jimeng web page after the user selects the mode and uploads the materials.
opencli commands
| Subcommand | Purpose |
|---|---|
generate-image2video <prompt> --image <path> |
Single reference image + motion prompt (type=video） |
history |
Verify after the task is complete |
new / workspaces |
Session management |
First-last frame, multiple storyboards, and multimodal still require manual mode selection on the web page; opencli has no --first/--last route.
Web page modes aligned with the prompt skill
| User input | Web side (user operation) | Prompt skill key points |
|---|---|---|
| Single image | Image-to-video, upload 1 | Write only the motion increment |
| First and last 2 images | First-last frame / transition | Describe the transition process |
| Multiple storyboard images | Multiple frames / storyboard | Narrative + transitions |
| Image + audio/video | All-in-one reference / multimodal | Composition relationship description |
Routing is done on theweb page; the agent uses the prompt skill to assist with the copy and uses opencli for sessions and acceptance.
Core process
1. PROMPT → jimeng-prompt-image2video
2. SESSION → 可选 opencli jimeng new --type=video
3. GEN → opencli jimeng generate-image2video "<motion prompt>" --image ./frame.png --wait 180
4. VERIFY → opencli jimeng history --limit N --type=video（视频可长达十余分钟，分轮查询）
Allowed opencli examples
opencli jimeng generate-image2video "镜头缓慢推进，发丝随风飘动" --image ./portrait.jpg --wait 180
opencli jimeng history --limit 10 --type=video
Prohibitions
- Must not call any dreamina video subcommand or
query_resultpolling. - Must not fabricate opencli image-to-video parameters.
- Must not guide the installation of dreamina.
And jimeng-cli-image2video
Users who have the dreamina CLI enabled use jimeng-cli-image2video for subcommand routing and polling; users of this skillmust notswitch to it.
Agent specifications
- The prompt describes only the part thatmoves, and does not repeat the static content of the image.
- Round-by-round
historyacceptance, shell infinite loops are prohibited. - For single-image image-to-video, use
generate-image2video --image; for complex modes, still guide the user to operate on the web page.
Gotchas
- First-last frame / storyboard / multimodal must be manual on the web page, and opencli parameters must not be fabricated.
- Video takes a long time, the default is
--wait=120, and 180+ is recommended for image-to-video. - For regular members, queuing and credits are subject to the web page.