Dumpling AI Extract Video
100 credits/call
ae_dumplingai_extract_video
Extract transcripts, frames, and insights from videos using DumplingAI.
100 credits/call
Pricing
100 credits/call
| Condition | Charge |
|---|---|
| Each call | 100 credits/call |
Parameters
Optional parameters use the platform defaults when omitted.
| Parameter | Type | Required | Description | Default |
|---|---|---|---|---|
video | string | Yes | 必填,待分析的视频媒体。inputMethod=url 时,传可由服务端直接下载并读取的视频文件 URL;不能把社交分享页面、播放器网页、登录页或 HTML 当作视频文件。分享链接须先用对应平台的视频获取工具或合法方式取得真实媒体,再分析。仅 HTTP 200 或 .mp4 后缀不能证明有效;签名直链可能无扩展名,不应仅按后缀拒绝。避免已过期、403 或要求登录/Cookie/Referer 才可读取的链接;本工具没有自定义下载鉴权参数。inputMethod=base64 时,传实际视频文件字节的 base64,不传网页 HTML、路径、URL 或编造数据。出现 ffprobe/moov atom/Invalid data 错误时,先检查是否下载到真实、完整、可解码的视频,不反复调用同一失败输入。 | — |
prompt | string | Yes | 必填,说明希望从视频提取或分析的内容,例如主要事件、口播要点或画面信息。不要把视频 URL/base64 放进 prompt 来代替必填 video。只依据实际可读取的视频回答;不臆造未能读取的内容。jsonMode=true 时,在此说明需要的 JSON 字段。 | — |
jsonMode | boolean | No | When true, requests the model to respond with JSON-formatted output. | false |
inputMethod | enum (url | base64) | Yes | 必填,视频文件内容的传入方式:url 表示可直接读取的视频媒体 URL;base64 表示视频文件字节的 base64 编码。两种方式须与 video 内容匹配。此字段不是社交网页解析模式,不要用它表示抖音/TikTok/微信分享页面。 | — |
requestSource | enum (API | WEB | MAKE_DOT_COM | ZAPIER | N8N | PLAYGROUND | DEFAULT_AUTOMATION | AGENT_PREVIEW | AGENT_LIVE | AUTOPILOT | STUDIO) | No | Optional identifier describing where the API request originated. | — |
Invocation example
Copy the request and explore its response example.
POST
/agent-api/v1/tool/executeX-Api-KeyRequest coderequest.sh
curl --request POST \
--url 'https://agentearth.ai/agent-api/v1/tool/execute?tool_name=ae_dumplingai_extract_video' \
--header "X-Api-Key: $AGENTEARTH_API_KEY" \
--header 'Content-Type: application/json' \
--data-raw '{
"params": {
"inputMethod": "url",
"jsonMode": true,
"prompt": "Analyze this Chinese e-commerce product video thoroughly. Extract the following information as structured JSON: 1) video_duration_seconds, 2) video_resolution_width and height, 3) fps, 4) ALL Chinese text visible in the video with timestamps (exact characters, where they appear, their approximate time ranges), 5) all human characters in the video (number, gender, age range, appearance, clothing, what actions they perform), 6) the product being shown and its features, 7) scene descriptions for each major shot/segment, 8) any text animations or motion effects observed. Be extremely thorough - identify every single Chinese character visible at any point in the video.",
"video": "https://file.gw.xnetwk.net/file/cf4b995d-d1dd-4fb7-bd57-d5ef3558a59b.mp4"
}
}'Response exampleresponse.json
{
"creditUsage": 670,
"prompt": "Analyze this Chinese e-commerce product video thoroughly. Extract the following information as structured JSON: 1) video_duration_seconds, 2) video_resolution_width and height, 3) fps, 4) ALL Chinese text visible in the video with timestamps (exact charact",
"results": [
{
"chinese_text": [
{
"location": "Top left",
"text": "腰酸",
"timestamp": "00:00 - 00:04"
}
],
"human_characters": [
{
"actions": [
"Leaning over with back pain while vacuuming"
],
"age_range": "20-30",
"appearance": "Slender build, long dark hair styled in a bun, light makeup, graceful movements.",
"character_count": 1,
"clothing": [
"Cream-colored summer dress (intro cleaning scene)"
],
"gender": "Female"
}
],
"motion_effects": [
"Text fades and slides in smoothly."
],
"product_details": {
"brand": "UB (优背)",
"color": "Soft Pink / Beige",
"key_features": [
"Double Y-shaped wide shoulder straps for pressure distribution"
],
"product_name": "Youbei (UB) Cloud-Sensing Posture Correction Garment"
},
"scene_descriptions": [
{
"description": "Vertical three-way split screen showing the woman experiencing back pain, hunching while working, and unstable core during yoga. Labels highlight the issues.",
"shot": "00:00 - 00:04"
}
],
"video_metadata": {
"aspect_ratio": "3:4",
"fps": 30,
"video_duration_seconds": 57,
"video_resolution": "1080x1440"
}
}
],
"videoDuration": 56.043
}This example shows part of the response. Copied content is valid JSON.

