Opennan

Copy a video

已上线

Send a video link. Get its shots with start and end times, the action, the camera move, the dialogue and an opening frame for each, plus the characters in it. Paid per started minute of video.

$1.223 per minuteFailed runs are not charged.

Copy a video

输入

  • night-race-input-c33d9d94.mp4Reading the length…

$1.223 per started minute.

These tools need an API key with the api:opennan scope: paste one under Use API key to run them.

输出

示例
示例输出Made with these inputs

4 shots

Length 0:08.00 · No watermark

  1. Shot 11

    0:00.00 – 0:01.96

    A black classic muscle car and a grey modern sports car sit side-by-side at a red traffic light on a wet, neon-lit city street at night, both stationary with brake lights glowing.

    Camera: static low-angle rear shot

  2. Shot 22

    0:01.96 – 0:06.42

    A dark-haired male driver stares forward with intense, focused eyes reflected in the rain-streaked rearview mirror, his sweaty face framed by blurred neon bokeh behind the wet windshield.

    Camera: static extreme close-up, rearview mirror

  3. Shot 33

    0:06.42 – 0:07.62

    A leather-gloved hand grips a spherical gear shifter and pushes it forward through a gear change, the wrist rotating slightly as the shift completes inside the dark cabin.

    Camera: static close-up, gear shifter

  4. Shot 44

    0:07.62 – 0:08.00

    A leather-gloved hand holds the gear shifter steady in its shifted position inside the dark car interior, the red-trimmed jacket sleeve visible and the cabin otherwise still.

    Camera: static close-up, gear shifter

1 characters

  • Character 1 · A man with dark hair, thick eyebrows, and a sweaty, intense face, seen driving a car at night.

运作方式

  1. 01Send the videoAn https link, trimmed to the part you need.
  2. 02We break it downEvery cut found, every shot described and checked.
  3. 03Get the breakdownShots, frames and characters, ready to recreate.

从代码调用 Copy a video

One endpoint and one key, on the same balance as the playground. The request below is the example's own.

POSThttps://api.opennan.com/v1/copy-video/detections
  1. 获取 API 密钥

    在控制台创建密钥。它与试用区使用同一余额。

    创建密钥
  2. 发送第一个请求

    与试用区发送的请求相同。把 OPENNAN_KEY 设为你的密钥。

    curl https://api.opennan.com/v1/copy-video/detections \
      -H "Authorization: Bearer $OPENNAN_KEY" \
      -H "Content-Type: application/json" \
      -d '{"video_url":"https://cdn.opennan.com/examples/copy-video-detect/night-race-input-c33d9d94.mp4"}'
    
    # Then poll until "status" is "succeeded" or "failed":
    curl https://api.opennan.com/v1/tasks/<task id> -H "Authorization: Bearer $OPENNAN_KEY"
  3. 查询任务

    调用会立即返回任务 ID。查询任务直到完成:

    GEThttps://api.opennan.com/v1/tasks/{id}
    queued
    等待处理
    running
    生成中
    succeeded
    完成:outputs[].url 中是文件
    failed
    失败:不收费

错误及解决方法

  • 401缺少密钥或密钥错误

    用 Authorization: Bearer 请求头发送控制台中的密钥。

  • 402余额不足

    在账单页面充值,然后重新发送同一请求。

  • 400有值不被接受

    查看 error.message,它会指出字段。请使用参数表中的值。

  • 429请求过多

    等几秒再试,每次遇到 429 都等得更久一些。

把 Opennan 添加到你的 AI 应用

安装一次,你的 AI 就能使用 Opennan 的全部工具:视频、图片、聊天、语音、音乐、音效、3D、字幕和配音。选择你的应用,按步骤操作。

MCP 服务器https://api.opennan.com/v1/mcp

适用于 Codex CLI、Codex 应用和 Codex IDE 扩展:它们共用 ~/.codex/config.toml。如果应用或 IDE 读不到 OPENNAN_KEY,请用写入密钥的 config.toml 方式。Codex 每次调用工具前会询问,批准即可继续。

  1. 获取 API 密钥

    在控制台创建密钥。密钥以 sk- 开头,与网站共用同一余额。

    创建密钥
  2. 把密钥保存为 OPENNAN_KEY

    # macOS / Linux (zsh shown; bash: ~/.bashrc)
    echo 'export OPENNAN_KEY="sk-..."' >> ~/.zshrc
    source ~/.zshrc
    
    # Windows (PowerShell), then open a new terminal
    setx OPENNAN_KEY "sk-..."
  3. 安装 Opennan 插件:包含服务器,以及教 AI 选择模型、确认价格、等待结果的技能

    codex plugin marketplace add Opennan/opennan-ai-plugin
    codex plugin add opennan@opennan
  4. 或者只添加服务器

    codex mcp add opennan --url https://api.opennan.com/v1/mcp --bearer-token-env-var OPENNAN_KEY
  5. 或者把这段加到 ~/.codex/config.toml

    [mcp_servers.opennan]
    url = "https://api.opennan.com/v1/mcp"
    http_headers = { Authorization = "Bearer sk-..." }
  6. 确认列表里有 opennan

    codex mcp list
  7. 重启 Codex,然后让它做任何事

    “做一段 5 秒的霓虹街道下雨视频,再配一首歌和一张海报图。”

把 sk-... 换成你自己的密钥并妥善保管:拿到它的人可以花你的余额。没有密钥时应用也能列出工具,但生成内容需要密钥。

在 AI 智能体中使用

把 Opennan MCP 服务器加到你的智能体中,然后调用此工具。

工具
detect_copy_video_shots
示例参数
{
  "video_url": "https://cdn.opennan.com/examples/copy-video-detect/night-race-input-c33d9d94.mp4"
}

Copy a video:常见问题

Copy a video 多少钱?

$1.223 per minute. The Run button shows the exact price for your input before you start.

如何从代码调用?

带上 API 密钥发送 POST https://api.opennan.com/v1/copy-video/detections。返回任务 ID:查询 https://api.opennan.com/v1/tasks/{id},直到任务成功或失败。

我的 AI 智能体能用吗?

能。添加 MCP 服务器 https://api.opennan.com/v1/mcp 并调用 detect_copy_video_shots 工具,或让智能体阅读 https://api.opennan.com/llms.txt。

运行失败会怎样?

失败的运行不收费。只为结果付费。

Which API key do I need?

A key with the api:opennan scope. A full api key does not include it: create one with that scope in the console, then use it here under Use API key or in your code.

我们只使用网站必需的 Cookie,例如你的语言和推荐码。没有广告。我们使用不依赖 Cookie 的分析工具统计页面访问量。详情请参阅我们的 Cookie 政策 。