数字人视频生成API说明

概述

魔珐星云数字人视频生成功能能够将文本、PPT转化为高质量数字人视频，助力开发者快速构建专业级视频生成能力。

API说明

1.鉴权

1.1 X-TOKEN计算

接口接收参数：

1.数据体：data = {"xxxx": "xxxx"} or {}

2.分配给外部的密钥：secret = "iamsecret"

3.接口方法路径（不包含host）：api_path = "/xxx/xxx?xx=xxx"

计算步骤：

1.将api_path转换为全小写形式：lower_api_path

2.将请求method方法转为小写形式：lower_method（例如:"delete"/"post"/...）

3.将data转换为json的字符串形式：sort_json_str

以python为例：json.dumps(dict(data), sort_keys=True).replace(' ', '')

4.按照如下顺序连接字符串：lower_api_path + lower_method + sort_json_str + secret + X-TIMESTAM

a.X-TIMESTAMP：接口秒级时间戳，距当前时间60s内有效

b.得到sign：/xxx/xxx?xx=xxx{"xxxx": "xxxx"}iamsecret1489133053

5.将sign以utf8编码，计算md5得到X-TOKEN：ddc6457fd0b373475ac65912b797ef05

1.2 接口调用

请求接口时应该加入如下头部信息：

• X-APP-ID：应用AK
• X-TIMESTAMP：秒级时间戳
• X-TOKEN：签名计算结果

1.3 demo代码

import time
import json
import hashlib
import requests
from urllib.parse import urljoin


def encode_with_md5(s):
    m = hashlib.md5()
    m.update(s.encode('utf-8'))
    return m.hexdigest()


def headers_need_sign(ak, secret, method, url, data):
    headers = {}
    t = int(time.time())

    data = json.dumps(dict(data), sort_keys=True).replace(' ', '')
    ori_sign = '{0}{1}{2}{3}{4}'.format(url.lower(), method.lower(), data, secret, t)
    sign = encode_with_md5(ori_sign)
    headers["X-APP-ID"] = ak
    headers["X-TOKEN"] = sign
    headers["X-TIMESTAMP"] = str(t)
    return headers


if __name__ == '__main__':
    ak = '37514ac0-3fce-4f4c-bc3f-86eba37da7dd'
    secret = 'bb81b706-ef1f-443e-9e86-9df8399f796b'
    method = 'POST'
    host = 'https://nebula-agent.xingyun3d.com'
    url = '/xxx/xxx?x=xx&z=zz'
    req_data = {
        "data1": "data1",
        "data2": "data2",
    }

    # 计算获取请求headers
    req_headers = headers_need_sign(ak, secret, method, url, req_data)

    # 请求接口
    req_url = urljoin(host, url)
    resp = requests.request(method, req_url, json=req_data, headers=req_headers)

2. API调用

2.1 发起渲染

方式1：通过segment发起渲染

host: https://nebula-agent.xingyun3d.com

请求路径：Post: /user/v1/video_synthesis_task/create_render_task

请求参数

返回参数：

方式2：通过PPT发起渲染

先调用 PPT解析接口 再调用 创建渲染任务接口

PPT解析接口：host: https://nebula-agent.xingyun3d.com

请求路径：Post: /user/v1/video_synthesis_task/parse_ppt_file

请求参数

返回参数

创建渲染任务接口：

host: https://nebula-agent.xingyun3d.com

请求路径：Post: /user/v1/video_synthesis_task/create_render_task

请求参数

返回参数

2.2查询视频结果

host: https://nebula-agent.xingyun3d.com

请求路径：GET: /user/v1/video_synthesis_task/get_render_task

请求参数

返回参数

2.3 取消渲染任务

host: https://nebula-agent.xingyun3d.com

请求路径：Post: /user/v1/video_synthesis_task/cancel_render_task

请求参数

返回参数

2.4 预览

host: https://nebula-agent.xingyun3d.com

请求路径：GET: /user/v1/video_synthesis_task/get_render_task_preview_url

请求参数

返回参数

3.错误码

DEMO示例

1. 发起渲染

1.1 方式1：通过segment发起渲染

Post: /user/v1/video_synthesis_task/create_render_task
Body: 
{
   "look_name": "AM054_Jap_9021_new",   
   "tts_vcn_name": "XMOV_HN_TTS__236",
   "studio_name": "youling_2d_v",
   "segment": [{
          "text": "为推动我国制造业数字化转型，促进智能制造技术正确应用，加强企业间友好交流合作，",
          "media_url": "https://media.xmov.ai/youyan/user_upload_qa/74275_4b733edeca68484e8940da.png"
       },
       {
          "text": "SIA上海国际工业自动化及机器人展览。整合多年行业资源、汇聚发展优势，将于2024年7月24至26日，在上海国家会展中心隆重举办。",
          "media_url": "https://media.xmov.ai/youyan/user_upload_qa/74275_df35e07846584dd98974bf.mp4"
       },
       {
          "text": "同期将举行多场论坛。邀请权威专家。知名企业。和主流厂商。分享智能制造发展的前沿趋势。典型案例。应用热点。以及实施策略。",
          "media_url": "https://media.xmov.ai/youyan/user_upload_qa/74275_1ca7a4a5407c43acb0da8b.mp4"
       }
    ],     # 可传 segment 或 PPT文件 
   # PPT文件 
   "ppt_file": 复古艺术.pptx    # binary
}   
Response:
{
    "error_code": 0,
    "error_reason": "",
    "data": {
        "task_id": 135
    }
}

1.2 方式2：通过PPT发起渲染

# coding=utf-8

import hashlib
import json
import time
import requests


def generate_x_token(data, secret, api_path, method, timestamp=None):
    lower_api_path = api_path.lower()
    lower_method = method.lower()
    sort_json_str = json.dumps(dict(data), sort_keys=True).replace(' ', '')
    x_timestamp = str(int(timestamp or time.time()))
    sign_str = f"{lower_api_path}{lower_method}{sort_json_str}{secret}{x_timestamp}"
    return hashlib.md5(sign_str.encode('utf-8')).hexdigest()


app_id = 'xxxxx'    # 替换成应用的App ID
secret = 'xxxxx'    # 替换成应用的App Secret
host = 'https://nebula-agent.xingyun3d.com'

# PPT解析接口
files = [
    ('ppt_file', ('PPT模板.pptx', open('PPT模板.pptx', 'rb'),
                  'application/vnd.openxmlformats-officedocument.presentationml.presentation'))
]
data= {}
api_path = '/user/v1/video_synthesis_task/parse_ppt_file'
method = 'POST'
timestamp = int(time.time())
headers = {}
headers["X-APP-ID"] = app_id
headers["X-TOKEN"] = generate_x_token(data, secret, api_path, method, timestamp)
headers["X-TIMESTAMP"] = str(timestamp)
start_time = time.time()
resp = requests.request(method, f'{host}/{api_path}', data=data,
                        headers=headers, timeout=30, files=files)
res = resp.json()
print(res)
parse_ppt_file_name = res.get('data').get('parse_ppt_file_name')


# 创建渲染任务接口
data = {"look_name": "caihongwei_3663_new", "tts_vcn_name": "XMOV_EN_TTS__13", "if_aigc_mark": True,
        "studio_name": "telestudio_simple_red_01", "sub_title": "on", "video_name": "测试调用API生成视频",
        "parse_ppt_file_name": parse_ppt_file_name}

api_path = '/user/v1/video_synthesis_task/create_render_task'
method = 'POST'
timestamp = int(time.time())
headers = {"content-type": "application/json"}
headers["X-APP-ID"] = app_id
headers["X-TOKEN"] = generate_x_token(data, secret, api_path, method, timestamp)
headers["X-TIMESTAMP"] = str(timestamp)

resp = requests.request(method, f'{host}/{api_path}', json=data,
                        headers=headers, timeout=30)
res = resp.json()
print(resp.json())
task_id = res.get('data').get('task_id')
data = {
    "task_id": task_id
}

2. 获取视频结果

GET: /user/v1/video_synthesis_task/get_render_task
Params: task_id: 135            # 任务id
Response:
{
    "error_code": 0,
    "error_reason": "",
    "data": {
        "id": 135,
        "error_reason": "合成任务失败，请重新发起或联系客服处理",
        "create_time": "2025-07-17T11:30:08.033710+08:00",
        "update_time": "2025-07-17T11:30:54.131480+08:00",
        "enable": true,
        "name": "6b187ef0c68c4ef2b9fda5964a392067",
        "video_name": "20250717_11_30_07.833",
        "output_resolution": "540P",
        "look_name": "AM058_10518_new",
        "tts_vcn_name": "XMOV_HN_TTS__6",
        "studio_name": "bust_chic_art_museum_01",
        "sub_title": "on",
        "synth_start_time": null,
        "synth_finish_time": null,
        "synth_state": "error",
        "segment": [
            {
                "text": "这是一条测试数据。",
                "media_id": 123338,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_b37eb4385ec64bec9fd4be.png"
            },
            {
                "text": "测试数据片段1",
                "media_id": 123339,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_b3195096c8d64cdfaa3b80.png"
            },
            {
                "text": "测试数据片段2",
                "media_id": 123340,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_e30188384695418b831ce6.png"
            },
            {
                "text": "测试数据片段3",
                "media_id": 123341,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_0cef3d5134474e87ad10b1.png"
            },
            {
                "text": "测试数据片段4",
                "media_id": 123342,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_97acf13816a0480f9b76ae.png"
            },
            {
                "text": "测试数据片段5",
                "media_id": 123343,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_5cd15131918e487f8184a2.png"
            },
            {
                "text": "测试数据片段6",
                "media_id": 123344,
                "media_url": "https://media.xmov.ai/youyan/user_upload_qa/171998_da32d7eece66455182f6ce.png"
            }
        ],
        "render_image_oss": null,
        "render_video_oss": null
    }
}

3. 取消渲染任务

Post: /user/v1/video_synthesis_task/cancel_render_task
{
   task_id: 135            # 任务id
}  
Response:
{
    "error_code": 0,
    "error_reason": ""
}