| schemaVersion | 2026-04-11T00:00:00.000Z |
| skillId | design/jimeng |
| name | jimeng |
| displayName | Jimeng |
| description | Generate images via Jimeng (即梦) AI image generation API. |
| aliases | ["jimeng","智能助手"] |
| version | 0.1.0 |
| sourceHash | sha256:4321838a05911bdac874f0585db8aec99f29a4cc0188f21c2957e9ca15dac21d |
| domain | design |
| departmentTags | ["ai-platform"] |
| sceneTags | ["content-generation","design"] |
Jimeng (即梦) Image Generation
Use the bundled script to generate images via the Jimeng AI image generation API.
Generate
uv run {baseDir}/scripts/generate_image.py --prompt "一只可爱的猫咪在窗台上晒太阳" --filename "output.png"
With model / resolution / aspect-ratio options
uv run {baseDir}/scripts/generate_image.py --prompt "赛博朋克风格的城市夜景" --filename "city.png" --model jimeng-5.0 --resolution 2k --aspect-ratio 16:9
Available models
jimeng-5.0 (latest)
jimeng-4.6
jimeng-4.5
jimeng-4.1
jimeng-4.0 (default)
jimeng-3.1
jimeng-3.0
Cookie / Session
JIMENG_COOKIES env var — one or more session IDs or full cookie strings, separated by ||| or ,
- Or set
skills.jimeng.env.JIMENG_COOKIES in ~/.openclaw/openclaw.json
Notes
- Resolutions:
1k (default), 2k.
- Aspect ratios:
1:1 (default), 4:3, 3:4, 16:9, 9:16.
- Use timestamps in filenames:
yyyy-mm-dd-hh-mm-ss-name.png.
- The script prints a
MEDIA: line for OpenClaw to auto-attach on supported chat providers.
- Do not read the image back; report the saved path only.