Shengsuanyun

https://docs.router.shengsuanyun.com
Admin
Shengsuanyun is an AI API platform that provides unified access to major language models and...
Tokens:569,418
Snippets:2,338
Trust Score:7
Update:1 month ago
Show doc for...
Tokens:
Raw
### Multi-frame Image-to-Video Example

Source: https://docs.router.shengsuanyun.com/358454722e0

Example request using start and end frames to control the generation process.

```yaml
              image-to-video-array:
                value:
                  model: runway/gen3a_turbo
                  images:
                    - uri: https://example.com/start.jpg
                      position: first
                    - uri: https://example.com/end.jpg
                      position: last
                  prompt: 从第一帧平滑过渡到最后一帧
                  ratio: '1280:768'
                  duration: 5
                  seed: 42
                summary: 多帧图生视频
                description: 使用首尾帧控制生成过程
```

--------------------------------

### Run Hermes Model Configuration

Source: https://docs.router.shengsuanyun.com/8561158m0

Initiate the interactive setup wizard for configuring model providers, including custom endpoints like Shengsuanyun. This command can be run after installation or manually.

```bash
hermes model
```

```bash
hermes setup
```

--------------------------------

### Video-to-Video with References Example

Source: https://docs.router.shengsuanyun.com/358454722e0

Example request for video-to-video generation including style references.

```yaml
              video-to-video-with-references:
                value:
                  model: runway/gen4_aleph
                  videoUri: https://example.com/input.mp4
                  prompt: 将视频转换为油画风格
                  ratio: '1280:720'
                  seed: 123
                  references:
                    - type: image
```

--------------------------------

### Basic Image-to-Video Example

Source: https://docs.router.shengsuanyun.com/358454722e0

Example request for generating a video from a single image.

```yaml
            examples:
              image-to-video-basic:
                value:
                  model: runway/gen4_turbo
                  image: https://example.com/image.jpg
                  prompt: 让图片中的场景动起来，添加微风和光影效果
                  ratio: '1280:720'
                  duration: 10
                summary: 基础图生视频
                description: 使用单张图片生成视频
```

--------------------------------

### Install n8n using npm

Source: https://docs.router.shengsuanyun.com/7989401m0

Use this command to install the n8n platform globally via npm. This is the recommended method for installation.

```bash
npm install n8n -g
```

--------------------------------

### Example CLI parameters

Source: https://docs.router.shengsuanyun.com/462296502e0

An example of setting the subdivision level and file format using command-line style parameters.

```text
'--subdivisionlevel medium --fileformat glb'
```

--------------------------------

### Basic Video-to-Video Example

Source: https://docs.router.shengsuanyun.com/358454722e0

Example request for converting an input video into a new style.

```yaml
              video-to-video-basic:
                value:
                  model: runway/gen4_aleph
                  videoUri: https://example.com/input.mp4
                  prompt: 将视频转换为卡通风格，保持原有动作
                  ratio: '1280:720'
                summary: 基础视频到视频
                description: 基于输入视频生成新风格视频
```

--------------------------------

### Base64 Image-to-Video Example

Source: https://docs.router.shengsuanyun.com/358454722e0

Example request using a Base64 encoded image string.

```yaml
              image-to-video-base64:
                value:
                  model: runway/gen4_turbo
                  image: >-
                    data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAUAAAAFCAYAAACNbyblAAAAHElEQVQI12P4//8/w38GIAXDIBKE0DHxgljNBAAO9TXL0Y4OHwAAAABJRU5ErkJggg==
                  prompt: 让场景充满生机
                  ratio: '960:960'
                  duration: 10
                summary: Base64 图生视频
                description: 使用 Base64 编码的图片生成视频
```

--------------------------------

### Install n8n-nodes-shengsuanyun via command line

Source: https://docs.router.shengsuanyun.com/7989401m0

Install the Shengsuanyun community node for n8n by running this npm command. Ensure n8n is stopped before installation and restarted afterward.

```bash
# 停止n8n服务
# 然后运行：
npm i n8n-nodes-shengsuanyun
# 重新启动n8n
n8n start
```

--------------------------------

### Content Moderation Control Example

Source: https://docs.router.shengsuanyun.com/358454722e0

Example request demonstrating how to configure content moderation thresholds.

```yaml
              with-content-moderation:
                value:
                  model: runway/gen4_turbo
                  image: https://example.com/portrait.jpg
                  prompt: 让人物自然地微笑
                  ratio: '1280:720'
                  duration: 10
                  contentModeration:
                    publicFigureThreshold: low
                summary: 内容审核控制
                description: 使用较低的公众人物识别阈值
```

--------------------------------

### Asynchronous Task Request Example (Raw JSON)

Source: https://docs.router.shengsuanyun.com/356147219e0

A raw JSON example showing the required fields for an asynchronous audio transcription request.

```json
                  file: >-
                    https://example-bucket.oss-cn-beijing.aliyuncs.com/audio/input-01.mp3
                  model: whisper-1
                  response_format: json
                  language: en
                summary: 异步任务请求示例（raw JSON）
```

--------------------------------

### Basic Chat Completion Request

Source: https://docs.router.shengsuanyun.com/chat/completions

A standard request example for chat completions using the deepseek-v3 model with streaming enabled.

```yaml
model: deepseek/deepseek-v3
stream: true
messages:
  - role: user
    content: 你好
stream_options:
  include_usage: true
```

--------------------------------

### Chat Completion Chunk Response

Source: https://docs.router.shengsuanyun.com/chat/completions

Example of a chat completion chunk containing search results and metadata.

```json
{
                                            "url": "https://new.qq.com/rain/a/20250411A01G8E00",
                                            "site_name": "腾讯网",
                                            "title": "上海天气今晚剧变!减少美国影片进口!男子睡觉1小时智驾狂飙百公里!女子花60万断骨增高!",
                                            "snippet": "今天(4月11日)天气:      【晨雾】早晨沿江沿海地区会出现能见度小于500米的浓雾,出行留意哦｡      【强对流天气】白天多云,最高气温29℃左右,最低气温15℃左右｡夜间到明天(4月12日)上午有强对流天气,伴有雷雨大风,雨量中到大雨,西部和北部大雨｡      未来天气:      【周末降温】明天上午,冷空气开始影响本市,白天气温一路下行,体感明显变凉变冷,周日早间最低气温降至10℃左右｡      【大风】周六和周日上午风力较大,陆地最大阵风7-8级,沿江沿海地区和长江口区8-10级,洋山港区和上海市沿海海面9-11级｡周日下午开始,风力逐渐减小｡"
                                        },
                                        {
                                            "url": "https://www.163.com/v/video/VRRTKO2VJ.html",
                                            "site_name": "网易",
                                            "title": "今日上海天气:15-29,多云转阴有阵雨或雷雨骤雨天气情况天气现象降温天气雷暴天气预报上海市_网易视频",
                                            "snippet": "点击按住拖动小窗 关闭 1997-2025 网易公司版权所有 About NetEase 不良信息举报 Complaint "
                                        }
                                    ]
                                },
                                "logprobs": null,
                                "finish_reason": null,
                                "index": 0
                            }
                        ]
                    }
```

--------------------------------

### Python Live API Integration Example

Source: https://docs.router.shengsuanyun.com/7867831m0

This example demonstrates how to initialize the `google-genai` client with a custom Base URL, establish a Live connection, send real-time audio input, and receive/save audio and text responses.

```APIDOC
## Python Live API Integration Example

### Description
This Python script shows how to use the `google-genai` SDK to connect to the Live API. It includes steps for configuring the client with a custom `BASE_URL`, sending local audio file data, and processing real-time audio and text responses.

### Configuration
- **API_KEY**: Your unique API key for authentication.
- **BASE_URL**: The endpoint for the Live API (e.g., `https://router.shengsuanyun.com/api`). The SDK handles WebSocket upgrades.
- **MODEL**: The specific Gemini model to use (e.g., `google/gemini-2.5-flash-live`).
- **CONFIG**: System instructions and response modalities, including `response_modalities: ["AUDIO"]`.

### Audio Handling
- **Input**: The example loads a local WAV file (`16000.wav`), resamples it to 16000Hz, and sends it as PCM 16-bit raw audio data.
- **Output**: Received audio responses are saved to `audio_output.wav`, expecting PCM 16-bit raw audio at 24000Hz.

### Code Structure
1. Import necessary libraries (`asyncio`, `io`, `wave`, `librosa`, `soundfile`, `genai`, `types`).
2. Define `API_KEY`, `BASE_URL`, `MODEL`, and `CONFIG`.
3. Implement an `async` function `main()`:
   - Initialize `genai.Client` with `http_options.base_url`.
   - Use `client.aio.live.connect()` to establish a session.
   - Load and prepare audio input.
   - Send audio using `session.send_realtime_input()`.
   - Iterate through `session.receive()` to process audio and text responses.
   - Save received audio to a WAV file.
   - Handle potential exceptions.
4. Run the `main()` function using `asyncio.run()`.

### Important Notes
- Ensure your network environment supports WebSocket communication.
- Input audio format is typically 16000Hz, 16-bit PCM. Output audio is typically 24000Hz, 16-bit PCM.
```

--------------------------------

### Install n8n using Docker

Source: https://docs.router.shengsuanyun.com/7989401m0

Run n8n in a Docker container. This command starts a container named 'n8n' and maps port 5678 for access.

```bash
docker run -it --rm --name n8n -p 5678:5678 n8nio/n8n
```

--------------------------------

### Run OpenClaw Onboarding Wizard

Source: https://docs.router.shengsuanyun.com/8561142m0

Initiate the interactive onboarding wizard to configure models, channels, and Gateway settings for OpenClaw.

```bash
openclaw onboard
```

--------------------------------

### Verify Claude Code Installation

Source: https://docs.router.shengsuanyun.com/8391850m0

After installation, run this command in your terminal to verify that Claude Code has been installed correctly and is accessible.

```bash
claude --version
```

--------------------------------

### Install Claude Code on Windows CMD

Source: https://docs.router.shengsuanyun.com/8391850m0

Install Claude Code on Windows using the Command Prompt. This command downloads, executes, and then deletes the installation script.

```cmd
curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd
```

--------------------------------

### Install Hermes Agent

Source: https://docs.router.shengsuanyun.com/8561158m0

Use this script to install the latest version of Hermes Agent. It automatically registers the 'hermes' command. Verify the installation with 'hermes --version'.

```bash
# Recommended one-click installation script (automatically registers hermes command)
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
# Or install from GitHub
# git clone https://github.com/NousResearch/hermes-agent.git && cd hermes-agent && make install
```

```bash
hermes --version
```

--------------------------------

### Create OpenCode Config Directory and File (Mac/Linux)

Source: https://docs.router.shengsuanyun.com/8119220m0

Command to create the necessary directory and open the OpenCode configuration file using nano on Mac or Linux.

```bash
mkdir -p ~/.config/opencode
nano ~/.config/opencode/opencode.json
```

--------------------------------

### Install Claude Code on macOS/Linux/WSL

Source: https://docs.router.shengsuanyun.com/8391850m0

Use this command to install Claude Code via curl on macOS, Linux, or Windows Subsystem for Linux. Ensure you have curl installed.

```bash
curl -fsSL https://claude.ai/install.sh | bash
```

--------------------------------

### Install Google Gemini Python SDK

Source: https://docs.router.shengsuanyun.com/7867831m0

Install the necessary Python packages for interacting with the Google Gemini API, including libraries for audio processing.

```bash
pip install google-genai soundfile librosa
```

--------------------------------

### Python Example for Real-time Audio Processing with Gemini Live API

Source: https://docs.router.shengsuanyun.com/7867831m0

This script demonstrates connecting to the Gemini Live API using the official Python SDK, sending real-time audio input, and receiving/saving audio responses. Ensure your API Key, Base URL, and model name are correctly configured. The script handles audio loading, sending, and saving, with error handling included.

```python
import asyncio
import io
import wave
import librosa
import soundfile as sf
from google import genai
from google.genai import types

# ================= 配置区域 =================
# 1. 替换为您的 API Key
API_KEY = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"

# 2. 替换为本平台的 Base URL
# 注意：SDK 会自动处理 WebSocket 连接升级
BASE_URL = "https://router.shengsuanyun.com/api"

# 3. 指定模型名称
MODEL = "google/gemini-2.5-flash-live"
# ===========================================

# 配置系统指令和响应模态
CONFIG = {
    "response_modalities": ["AUDIO"],
    "system_instruction": "You are a helpful assistant and answer in a friendly tone."
}

async def main():
    # 初始化客户端，配置自定义 Base URL
    client = genai.Client(
        api_key=API_KEY,
        http_options=types.HttpOptions(
            base_url=BASE_URL,
        ),
    )

    try:
        print(f"正在连接到 {BASE_URL}...")

        # 建立 Live 连接
        async with client.aio.live.connect(model=MODEL, config=CONFIG) as session:
            print("连接成功！")

            # --- 准备音频输入 (示例：读取本地 wav 文件) ---
            print("正在加载音频文件...")
            # 这里假设您有一个名为 16000.wav 的文件，采样率为 16k
            # 实际使用中可以是麦克风实时流
            buffer = io.BytesIO()
            # 使用 librosa 加载并重采样为 16000Hz
            y, sr = librosa.load("16000.wav", sr=16000)
            sf.write(buffer, y, sr, format='RAW', subtype='PCM_16')
            buffer.seek(0)
            audio_bytes = buffer.read()
            print(f"已加载 {len(audio_bytes)} 字节音频数据")

            # --- 发送音频数据 ---
            print("正在发送音频...")
            await session.send_realtime_input(
                audio=types.Blob(data=audio_bytes, mime_type="audio/pcm;rate=16000")
            )
            print("音频发送完毕，等待响应...")

            # --- 接收并保存音频响应 ---
            # 输出文件设置：单声道, 16位, 24000Hz (Gemini Live 默认输出 24k)
            wf = wave.open("audio_output.wav", "wb")
            wf.setnchannels(1)
            wf.setsampwidth(2)
            wf.setframerate(24000)

            response_count = 0
            async for response in session.receive():
                # 处理音频数据
                if response.data is not None:
                    wf.writeframes(response.data)
                    response_count += 1
                    print(f"收到音频片段 #{response_count}")

                # 处理文本数据 (如果有)
                if hasattr(response, 'text') and response.text:
                    print(f"收到文本响应: {response.text}")

            wf.close()
            print(f"\n完成！共收到 {response_count} 个音频片段。\n")
            print("响应音频已保存至 audio_output.wav")

    except Exception as e:
        print(f"\n发生错误: {type(e).__name__}: {e}")
        import traceback
        traceback.print_exc()

if __name__ == "__main__":
    asyncio.run(main())

```

--------------------------------

### Configure config.toml on Mac/Linux

Source: https://docs.router.shengsuanyun.com/8751512m0

Use bash commands to create the config.toml file with Shengsuanyun provider details on Mac or Linux systems.

```bash
cat > ~/.codex/config.toml << 'EOF'
model = "anthropic/claude-sonnet-4.6"
model_provider = "shengsuanyun"

[model_providers.shengsuanyun]
name = "胜算云"
base_url = "https://router.shengsuanyun.com/api/v1"
env_key = "OPENAI_API_KEY"
wire_api = "responses"
EOF
```

--------------------------------

### v1TemplateSpecV1 Example Configuration

Source: https://docs.router.shengsuanyun.com/282986759d0

An example of a template specification for a text generation task, defining metadata, execution steps, input schemas, and field bindings.

```yaml
meta:
  name: 文案生成模板
  description: 根据提示词生成一句短文案
  primaryOutputType: text
steps:
  - stepId: stp_txt001
    displayName: 生成文案
    executionUnit: text-generate
    instruction: 根据用户输入生成一句简短中文营销文案。
    defaultModelRef:
      modelKey: google/gemini-2.5-flash
    allowModelOverride: true
inputSchema:
  fields:
    - key: prompt
      label: 提示词
      description: 希望模型参考的原始需求
      required: true
      valueType: string
      order: 1
      sourceKind: inline_text
    - key: text_model
      label: 文本生成模型（可选）
      description: 留空时使用默认模型
      required: false
      valueType: string
      order: 2
      sourceKind: model_id
  instructions:
    - 每行代表一次文案生成任务。
    - 模型列留空时使用模板默认模型。
  sampleRows:
    - values:
        prompt: 为一盏极简白色台灯写一句商品海报文案
        text_model: ''
fieldBindings:
  - fieldKey: prompt
    stepId: stp_txt001
    paramKey: prompt
    bindMode: replace
  - fieldKey: text_model
    stepId: stp_txt001
    paramKey: model
    bindMode: replace
```

--------------------------------

### Successful Response Example

Source: https://docs.router.shengsuanyun.com/353574243e0

Example of a successful API response containing image URLs and status metadata.

```json
{
  "code": 10000,
  "data": {
    "algorithm_base_resp": {
      "status_code": 0,
      "status_message": "success"
    },
    "binary_data_base64": [],
    "image_urls": [
      "https://p9-aiop-sign.byteimg.com/tos-cn-i-vuqhorh59i/20250919151238BFC19FD8953511507BE4-5111-0~tplv-vuqhorh59i-image-v1.image?rk3s=7f9e702d&x-expires=1758352362&x-signature=5Oa0eb7f7D6SKUQBeJyKnanNh9I%3D"
    ]
  },
  "message": "Success",
  "request_id": "20250919151238BFC19FD8953511507BE4",
  "status": 10000,
  "time_elapsed": "3.72549491s"
}
```
Tokens:
Raw