Generate per-slide avatar video (avatar or lipsync mode)

POST

/ai/learningPaths/generateVideo

Initiate an AI per-slide video render. Mode 'avatar' renders a fresh talking-head from the slide's existing audio and a provider-shaped avatarId; mode 'lipsync' rewrites a supplied source videoUrl's lip movements to match the slide's audio. One (slideId, lang) pair per request — the cross-language follow-up popup spawns N parallel requests rather than batching here. Queue-backed (AW-722): returns a durable jobId immediately; idempotent on re-submit, preempts a stale in-flight render, and locks the slide translation while running.

Request

Body Params application/jsonRequired

Examples

Responses

🟢200OK

application/json

Learning path video generation started

Bodyapplication/json

🟠401Unauthorized

Request Example

Shell

JavaScript

Java

Swift

curl --location 'https://api.aworld.cloud/dashboard/v1/ai/learningPaths/generateVideo' \
--header 'Content-Type: application/json' \
--data '{
    "learningPathId": "string",
    "slideId": "string",
    "lang": "en",
    "mode": "avatar",
    "avatarId": "string",
    "imageUrl": "http://example.com",
    "prompt": "string",
    "provider": "longcat"
}'

Response Example

{
    "jobId": "FUMzPf00hz_kcj2R2Hlzq",
    "status": "queued",
    "createdAt": "string"
}

Modified at 2026-07-22 08:08:42

Generate learning path audio narration

Translate learning path