生成 AI 影片
直接從你的後端

1. 1

   挑一個影片代理

   代理就是存好的模型設定：引擎、解析度、比例與預設值。列出你工作區裡的代理，記住 id。

   curlresponse

   複製

   export AUTOSTUD_API_KEY="sk_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
   
   # A video agent is a saved model configuration, not a model name.
   curl "https://autostud.ai/api/v1/agents?agent_type=video&limit=10" \
     -H "Authorization: Bearer $AUTOSTUD_API_KEY"

2. 2

   送出提示詞

   可以只給文字，也可以給一張要動起來的靜態圖。回應會帶回生成 id 和花掉的點數。

   curlimage to videoresponse

   複製

   curl -X POST https://autostud.ai/api/v1/actions/videos_ai.generate \
     -H "Authorization: Bearer $AUTOSTUD_API_KEY" \
     -H "Content-Type: application/json" \
     -H "Idempotency-Key: lighthouse-shot-01" \
     -d '{
       "agent_id": "b21f7d90-3c44-4e1a-9f77-0c2ad5e6b118",
       "prompt": "A lighthouse at dusk, slow push-in, waves breaking under it",
       "duration": 8
     }'

3. 3

   取回影片

   接 generation.completed 這個 webhook，或者你偏好輪詢就再讀一次生成紀錄。託管網址就在紀錄上。

   webhookpollingevent

   複製

   # Register once. Every generation reports itself from then on.
   curl -X POST https://autostud.ai/api/v1/webhooks \
     -H "Authorization: Bearer $AUTOSTUD_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
       "url": "https://your-app.com/hooks/autostud",
       "events": ["generation.completed", "generation.failed"],
       "description": "AI video"
     }'
   # The signing secret comes back once, in this response. Store it.

開發者最先問的問題