video-effect-apply

Video Effect Apply

Overview

YesAlready has video or imageApply overall visual style transformation or creative effects (without modifying/replacing/removing specific content elements), covering 3 scenarios + 2 fallbacks:

1. Preset style effects (video input): General default entry point, the Agent uses prompt to search the preset effect library for available effect_id。
2. AI creative effects (image input only)：image_url + without video_url, generate an AI creative video based on the image.
3. Long-video style transfer: overall style conversion such as anime/pixel/clay.

Not handled: video content replacement, video removal, video quality restoration, video canvas expansion, text-only video generation, video splicing.

API Mapping

Scenario	API name	Required input
Preset style effects (video)	formula_group_v2v_creative	video_url
AI creative effects (image)	formula_group_video_creative	image_url(no video)
Long-video style transfer	aiflow_long_video_redrawn	video_url
Fallback (v2v, video input)	video_molasses_edit	video_url
Fallback (video effects, image input)	video_bonbon_img2vid_v30	image_url

model mapping:auto(default, per routing rules) / molasses → video_molasses_edit / bonbon → video_bonbon_img2vid_v30。

Dependencies

• meitu-cli: >=2.0.6
• Credentials：CONFIG AKSK → meitu tools update; EXEC AKSK → run command
• Environment variables：MEITU_OPENAPI_TOOL_TASK_MODE=command

Path alias: $VISUAL = {OPENCLAW_HOME}/workspace/visual/

Core Workflow

Preflight → Execute → Deliver

Preflight

1. meitu --version and meitu auth verify --json
2. Detect MEITU_OPENAPI_TOOL_TASK_MODE=command
3. output_dir: project → ./output/; one-off → $VISUAL/output/video-effect-apply/；mkdir -p

Execute

Trigger signals and routing

Core judgment dimensions:Input type (image vs video) + prompt signal words → effect dispatch。

Decision order:

1. Input type judgment (routing iron rule)：

   • Image-only input (no video_url）→ force to formula_group_video_creative(fallback to video_bonbon_img2vid_v30）
   • Video input present → continue dispatching by prompt signal words

2. Video input routing：

   • The tool internally uses prompt to search the preset effect library
     • Available effect_id → formula_group_v2v_creative
     • not found → proceed to signal word judgment
   • Signal word judgment:
     • Anime / pixel / clay / style transfer → aiflow_long_video_redrawn
     • General / unclear → fallback video_molasses_edit

3 API routing quick reference

API	Required input	Trigger keywords
formula_group_v2v_creative	video_url	Add effects, preset style, creative玩法
formula_group_video_creative	image_url no video(iron rule)	AI creative video
aiflow_long_video_redrawn	video_url	Style transfer, anime, pixel, clay

Parameter definitions

Parameter	Type	Required	Range	Default	Description
video_url	STRING	Conditionally required	—	—	v2v_creative / long_video_redrawn required
image_url	STRING	Conditionally required	—	—	video_creative required
prompt	STRING	Yes	—	—	The user's desired effect/style description, based on which the tool internally retrieves matches. effect_id
model	STRING	No	auto/molasses/bonbon	auto	Model selection

Note:effect_id Not used as an external input parameter, automatically retrieved internally by the tool based on prompt .

Tool call

# 视频输入
meitu video-effect-apply --video_url {url} --prompt "{特效描述}" --json --download-dir {output_dir} --skill_name skill_video-effect-apply

# 图片输入（铁律）
meitu video-effect-apply --image_url {url} --prompt "{特效描述}" --json --download-dir {output_dir} --skill_name skill_video-effect-apply

# 显式指定 model
meitu video-effect-apply --video_url {url} --prompt "..." --model molasses --json --skill_name skill_video-effect-apply ...

Error fallback

Scenario	Handling
video_url and image_url are both missing	Prompt "Please provide an image or video"
"AI creative video" is requested but no image is provided	Prompt "Please provide an image"
"Preset effect" is requested but no video is provided	Prompt "Please provide a video link"
video_url / image_url is inaccessible	Do not retry, prompt "Please provide a valid link"
Video input retrieval effect_id no match	Fallback video_molasses_edit
Image input retrieval effect_id no match	Fallback video_bonbon_img2vid_v30
formula_group_v2v_creative failed	Retry once; if it still fails, use fallback video_molasses_edit
formula_group_video_creative failed	Retry once; if it still fails, use fallback video_bonbon_img2vid_v30
"Style transfer" but the video is <3s	Call normally and inform that the short video effect may be unsatisfactory
Any API failure / timeout	Retry once; if it still fails, return an error
Content compliance interception	Return the compliance prompt directly, do not retry

Deliver

mv {file} {output_dir}/{date}_{name}_effect-apply.mp4

Output

• Format：MP4
• Naming：{YYYY-MM-DD}_{descriptive-name}_effect-apply.mp4
• Location: Project → ./output/; one-time → $VISUAL/output/video-effect-apply/

Baseline Task ID

t_mt1a3i5n7ba09180a6-3aef-4a4d-8acf-53b2fbb033f4