AdsTurbo 虚拟人口播 · AdsTurbo Digital Human Voiceover

让虚拟人把文案念出来：可选平台数字人形象、用照片+声音克隆专属形象，或直接给一张人像加一段音频做对口型。含 TTS 配音与形象管理。

Install

openclaw skills install @adsturbo/adsturbo-digital-human

AdsTurbo Digital Human Voiceover

Have a virtual human avatar read a script aloud, producing a talking-head video.

When to use this skill

Use it when the user says "digital human", "voiceover", "have XX say this", "make a digital human from my photo", "lip sync", "dub", or "clone a voice".

To generate footage out of thin air (with no one speaking) use adsturbo-video-generation; to swap out the person in an existing video use adsturbo-video-transform.

Prerequisites

Python 3.8+ (the scripts use the standard library only; nothing to install)
An AdsTurbo API key in the environment variable ADSTURBO_API_KEY (see below for how to get one)
Optional ADSTURBO_BASE_URL, default https://adsturbo.ai/klian/novartapi

Getting an API key

Sign up or sign in at https://adsturbo.ai?channel=clawhub
Open https://adsturbo.ai/iris/api (Workspace → API), click "Create API Key" and copy the key
export ADSTURBO_API_KEY=<your key>

Every plan can call the API with an API key, and usage is billed in credits; if you run out, top up or upgrade at https://adsturbo.ai/pricing. 1080p+ video and 4K images need a subscription (credit packs alone don't count), and creating a custom digital human needs a Pro plan.

If the user has no key yet: when ADSTURBO_API_KEY is not set, or a script reports a missing key or a 401, do not keep retrying. First finish everything that doesn't need the API (clarify the brief, draft the script or prompt, settle aspect ratio, duration, model and other parameters, prepare the exact command), then tell the user they need an AdsTurbo API key and give them the three steps above; once they have it, run the prepared command. Without a key the scripts print "error": "api_key_required" and the steps on stdout. Keep the key in the environment variable and never echo it back in a reply.

First decide where the avatar comes from

This is the single fork point for picking a command:

text

Who does the user want speaking?
├─ Doesn't matter, pick a good-looking one → choose from the actors list → perform
├─ Use my/our own person, reused repeatedly → persona-create to register → perform
└─ Just this one image, one-off             → lipsync

bash

# Platform avatar
python3 scripts/digital_human.py actors --gender female --limit 10
python3 scripts/digital_human.py perform --actor-id act_123 --script "script content"

# Dedicated avatar (photo + optional voice cloning)
python3 scripts/digital_human.py persona-create --photo-url https://.../me.jpg --voice-audio-url https://.../voice.mp3
python3 scripts/digital_human.py perform --actor-id <returned actor_id> --script "script content"

# One-off lip sync
python3 scripts/digital_human.py lipsync --avatar-url https://.../face.jpg --audio-url https://.../voice.mp3

Full parameters in references/digital_human.md.

Hard constraint: assets must be public URLs only

None of the --photo-url, --avatar-url, --audio-url parameters accept local file paths. When the user provides a local file, upload it first:

bash

python3 scripts/upload.py image ./portrait.jpg
python3 scripts/upload.py audio ./voice.mp3

See references/upload.md for details.

Handling async tasks

perform / persona-create / lipsync are all asynchronous. The scripts poll automatically after submission until a result is ready, so under normal conditions you can just wait.

On timeout: use query --workspace-id <id> to keep waiting — do not resubmit. The task is still running, and resubmitting will double-charge.
To submit without waiting: add --no-wait
To check progress on a batch: python3 scripts/work.py batch --workspace-ids ws_a ws_b

See references/work.md for details.

When replying to the user

Give the result link directly — don't mention script names, command lines, JSON, or exit codes
After submission, tell the user roughly how long to expect: perform 2–5 minutes, persona-create 1–3 minutes, lipsync 2–5 minutes
On error, explain the problem in one sentence, then ask whether to retry
perform will fail before persona-create completes — check persona-status first in that case

Scope boundaries

This skill only handles "making a digital human speak". The following are out of scope — don't promise them to the user:

User wants	Use instead
Text-to-video / image-to-video	`adsturbo-video-generation`
Replicate an existing ad	`adsturbo-ad-clone`
Watermark removal / 4K / subtitles	`adsturbo-video-enhance`
Character swap / video translation / motion control	`adsturbo-video-transform`
Image generation / e-commerce images	`adsturbo-image`

AdsTurbo 虚拟人口播

让一个虚拟人形象把文案念出来，产出口播视频。

什么时候用这个 skill

用户说「数字人」「口播」「让 XX 说这段话」「用我的照片做个数字人」「对口型」「配音」「克隆声音」时用它。

要凭空生成画面（不涉及人说话）用 adsturbo-video-generation；要把已有视频里的人换掉用 adsturbo-video-transform。

前置条件

Python 3.8+（脚本只用标准库，不需要安装依赖）
AdsTurbo API Key，放在环境变量 ADSTURBO_API_KEY 里，获取方法见下
可选 ADSTURBO_BASE_URL，默认 https://adsturbo.ai/klian/novartapi

获取 API Key

在 https://adsturbo.ai?channel=clawhub 注册或登录
打开 https://adsturbo.ai/iris/api（工作台 → API），点「创建 API 密钥」，复制生成的 Key
export ADSTURBO_API_KEY=<你的 Key>

任何套餐都能用 API Key 调用，按用量扣积分；积分不够时到 https://adsturbo.ai/pricing 充值或升级。1080p 及以上视频、4K 图片需要订阅（只买积分包不算），创建自定义数字人需要 Pro 套餐。

用户还没有 Key 时：ADSTURBO_API_KEY 没设置、脚本报 Key 缺失或 401 时，不要反复重试。先把不需要调接口就能做的部分做完（理清需求，写好文案或提示词，定好画幅、时长、模型等参数，拼好要执行的完整命令），再告诉用户需要 AdsTurbo API Key，把上面三步发给他；拿到 Key 后直接执行准备好的命令。脚本缺 Key 时会在 stdout 输出 "error": "api_key_required" 和获取步骤。Key 只放在环境变量里，回复里不要回显。

先判断形象从哪来

这是选命令的唯一分岔点：

text

用户想让谁来说？
├─ 无所谓，挑个好看的        → actors 列表里选 → perform
├─ 用我/我们的人，要反复用   → persona-create 建号 → perform
└─ 就这张图，一次性          → lipsync

bash

# 平台形象
python3 scripts/digital_human.py actors --gender female --limit 10
python3 scripts/digital_human.py perform --actor-id act_123 --script "文案内容"

# 专属形象（照片 + 可选声音克隆）
python3 scripts/digital_human.py persona-create --photo-url https://.../me.jpg --voice-audio-url https://.../voice.mp3
python3 scripts/digital_human.py perform --actor-id <返回的 actor_id> --script "文案内容"

# 一次性对口型
python3 scripts/digital_human.py lipsync --avatar-url https://.../face.jpg --audio-url https://.../voice.mp3

完整参数见 references/digital_human.md。

硬约束：素材只收公网 URL

所有 --photo-url --avatar-url --audio-url 参数不接受本地文件路径。用户给本地文件时先上传：

bash

python3 scripts/upload.py image ./portrait.jpg
python3 scripts/upload.py audio ./voice.mp3

详见 references/upload.md。

异步任务怎么处理

perform / persona-create / lipsync 都是异步的，脚本默认提交后自动轮询到出结果，正常情况直接等即可。

超时了：用 query --workspace-id <id> 接着等，不要重新提交——任务还在跑，重交会重复扣费
只想提交不等：加 --no-wait
批量查进度：python3 scripts/work.py batch --workspace-ids ws_a ws_b

详见 references/work.md。

回复用户时

直接给结果链接，不要提脚本名、命令行、JSON 或退出码
提交后告诉用户大概要等多久：perform 2–5 分钟，persona-create 1–3 分钟，lipsync 2–5 分钟
出错时用一句话说清是什么问题，再问要不要重试
persona-create 完成前 perform 会失败，这时先查 persona-status

能力边界

这个 skill 只做「让虚拟人说话」。以下不在范围内，不要向用户承诺：

用户想要	该用
文生视频 / 图生视频	`adsturbo-video-generation`
照着一条广告复刻	`adsturbo-ad-clone`
去水印 / 4K / 字幕	`adsturbo-video-enhance`
换角色 / 视频翻译 / 动作控制	`adsturbo-video-transform`
生成图片 / 电商图	`adsturbo-image`