SiliconFlow

https://docs.siliconflow.com/en/
Admin
SiliconFlow is a cloud service platform providing developers with affordable, high-speed API access...
Tokens:109,394
Snippets:813
Trust Score:9.4
Update:2 weeks ago
Show doc for...
Tokens:
Raw
### Install oh-my-openagent via OpenCode prompt

Source: https://docs.siliconflow.com/en/usercases/how-to-build-a-personal-wiki-knowledge-base-andrej-karpathys-pattern-with-siliconflow-and-opencode

Paste this prompt into OpenCode to install and configure the oh-my-openagent plugin. The URL points to the official installation guide for manual setup, provider authentication, and troubleshooting.

```text
Install and configure oh-my-openagent by following the instructions here:
https://raw.githubusercontent.com/code-yeongyu/oh-my-openagent/refs/heads/dev/docs/guide/installation.md

```

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

### Install Hermes Agent

Source: https://docs.siliconflow.com/en/usercases/how-to-use-hermes-agent-with-siliconflow-apis

Installs Hermes Agent and automatically launches the Hermes Setup Wizard. Press Enter to continue after installation completes.

```bash
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
```

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

### Start Hermes Gateway Setup

Source: https://docs.siliconflow.com/en/usercases/how-to-deploy-an-ai-assistant-on-discord-using-hermes-agent

Run this command in your terminal to start the Hermes Agent gateway setup. When prompted, select Discord and paste your Bot Token and User ID.

```bash
hermes gateway setup

```

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

### Launch OpenCode

Source: https://docs.siliconflow.com/en/usercases/how-to-use-opencode-with-siliconflow-apis

Starts the OpenCode CLI in your terminal. Run this after installation to begin using OpenCode.

```bash
opencode

```

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

### Example query to ChatGPT using llms-full.txt

Source: https://docs.siliconflow.com/en/userguide/use-docs-with-cursor

Shows how to ask a question to ChatGPT using the llms-full.txt URL to get information about SiliconFlow.

```text
https://docs.siliconflow.com/llms-full.txt  What does SiliconFlow focus on?

```

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

### Run OpenClaw onboarding

Source: https://docs.siliconflow.com/en/usercases/how-to-use-openclaw-with-siliconflow-apis

Starts the OpenClaw onboarding process in the terminal. During onboarding, select 'Yes' to proceed, choose 'QuickStart', 'Use existing values', skip the model/auth provider step, select 'siliconflow' as the provider, keep the current default model, choose Discord (Bot API), paste the bot token, and configure skills.

```bash
openclaw onboard

```

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

### Install OpenCode via curl

Source: https://docs.siliconflow.com/en/usercases/how-to-use-opencode-with-siliconflow-apis

Installs OpenCode using the official installation script. Run this command in your terminal before launching OpenCode.

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

```

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

### Example JSON output

Source: https://docs.siliconflow.com/en/userguide/guides/json-mode

The model's output for the example request, showing the JSON object with the champions' names.

```json
{"Men's Champion": "Ma Long", "Women's Champion": "Chen Meng"}
```

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

### Successful response example

Source: https://docs.siliconflow.com/en/api-reference/userinfo/get-user-info

Example of a 200 response with user data including id, name, email, balance, and status.

```json
{
  "code": 20000,
  "message": "OK",
  "status": true,
  "data": {
    "id": "userid",
    "name": "username",
    "image": "user_avatar_image_url",
    "email": "user_email_address",
    "isAdmin": false,
    "balance": "0.88",
    "status": "normal",
    "introduction": "",
    "role": "",
    "chargeBalance": "88.00",
    "totalBalance": "88.88"
  }
}
```

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

### Install OpenClaw on Windows (PowerShell)

Source: https://docs.siliconflow.com/en/usercases/how-to-use-openclaw-with-siliconflow-apis

Install OpenClaw on Windows using PowerShell.

```powershell
iwr -useb https://openclaw.ai/install.ps1 | iex

```

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

### Successful response example

Source: https://docs.siliconflow.com/en/api-reference/systemone/create-systemone

Example of a 200 response with model, answers, and usage token counts.

```json
{
  "model": "Kev-4B",
  "answers": {
    "is_urgent": {
      "type": "noul",
      "noul": 0.92
    }
  },
  "usage": {
    "input_tokens": 296,
    "output_tokens": 20
  }
}
```

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

### Install DB-GPT Python Package

Source: https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-DB-GPT

Install the DB-GPT Python package and additional dependencies for testing.

```bash
pip install "dbgpt>=0.6.3rc2" openai requests numpy

```

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

### Start the DB-GPT service

Source: https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-DB-GPT

Starts the DB-GPT web server on port 5670. After running, access the deployed DB-GPT at http://127.0.0.1:5670/.

```bash
dbgpt start webserver --port 5670

```

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

### Example answers object in SystemOne response

Source: https://docs.siliconflow.com/en/api-reference/systemone/create-systemone

Shows a JSON example of the 'answers' field in the response. The answer map is keyed exactly like the request questions.

```json
{  
  "is_urgent": { "type": "noul", "noul": 0.92 }  
}
```

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

### GET /v1/models

Source: https://docs.siliconflow.com/en/api-reference/models/get-model-list

This snippet documents the GET /v1/models endpoint, including its query parameters, authentication, request example, and response schema.

```APIDOC
## GET /v1/models

### Description
Retrieves a list of available models from the SiliconFlow API.

### Method
GET

### Endpoint
https://api.siliconflow.com/v1/models

### Parameters
#### Query Parameters
- **type** (enum<string>) - Optional - The type of models. Available options: `text`, `image`, `audio`, `video`.
- **sub_type** (enum<string>) - Optional - The sub type of models. You can use it to filter models individually without setting type. Available options: `chat`, `embedding`, `reranker`, `text-to-image`, `image-to-image`, `speech-to-text`, `text-to-video`.

### Authorizations
- **Authorization** (string, header) - Required - Use the following format for authentication: `Bearer <token>`.

### Request Example
```
curl --request GET \
  --url https://api.siliconflow.com/v1/models \
  --header 'Authorization: Bearer <token>'
```

### Response
#### Success Response (200)
- **object** (string) - Example: `"list"`.
- **data** (object[]) - Array of model objects. Each model object contains:
  - **id** (string) - The model identifier.
  - **object** (string) - The object type, e.g., `"model"`.
  - **created** (integer) - Creation timestamp.
  - **owned_by** (string) - The owner of the model.

#### Response Example
```
{
  "object": "list",
  "data": [
    {
      "id": "stabilityai/stable-diffusion-xl-base-1.0",
      "object": "model",
      "created": 0,
      "owned_by": ""
    }
  ]
}
```

### Error Responses
- **400** - Bad Request
- **401** - Unauthorized
- **404** - Not Found
- **429** - Too Many Requests
- **503** - Service Unavailable
- **504** - Gateway Timeout
```

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

### Get Model List with cURL, Python, and JavaScript

Source: https://docs.siliconflow.com/en/api-reference/models/get-model-list

Shows the same GET /v1/models request in cURL, Python, and JavaScript. The endpoint returns a list of available models and requires a Bearer token in the Authorization header. The Python example uses the requests library; the JavaScript example uses fetch.

```bash
curl --request GET \
  --url https://api.siliconflow.com/v1/models \
  --header 'Authorization: Bearer <token>'
```

```python
import requests

url = "https://api.siliconflow.com/v1/models"

headers = {"Authorization": "Bearer <token>"}

response = requests.get(url, headers=headers)

print(response.text)
```

```javascript
const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.siliconflow.com/v1/models', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

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

### Example questions object for SystemOne request

Source: https://docs.siliconflow.com/en/api-reference/systemone/create-systemone

Shows a minimal JSON example for the 'questions' field in the request body. The key is chosen by the caller and the response uses the same key.

```json
{  
  "is_urgent": {  
    "type": "noul",  
    "instructions": "Does this message convey urgency?"  
  }  
}
```

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

### Verify Node.js version

Source: https://docs.siliconflow.com/en/usercases/how-to-use-openclaw-with-siliconflow-apis

Check that Node.js version 22 or newer is installed before proceeding with OpenClaw setup.

```bash
node --version

```

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

### Run Open Design from source

Source: https://docs.siliconflow.com/en/usercases/how-to-use-open-design-with-siliconflow-apis

Run Open Design from source. Requires Node ~24 and pnpm 10.33.x; starts the dev web server with pnpm tools-dev run web.

```bash
git clone https://github.com/nexu-io/open-design.git
cd open-design
corepack enable && pnpm install
pnpm tools-dev run web
```

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

### Technical Documentation Generation with Qwen/Qwen2.5-Coder-32B-Instruct

Source: https://docs.siliconflow.com/en/userguide/capabilities/text-generation

Generates a Python tutorial on asynchronous web scraping using the Qwen2.5-Coder-32B-Instruct model. Requires an API key and uses the SiliconFlow base URL. Sets temperature to 0.7 and max_tokens to 4096.

```python
from openai import OpenAI
client = OpenAI(api_key="YOUR_KEY", base_url="https://api.siliconflow.com/v1")
response = client.chat.completions.create(  
    model="Qwen/Qwen2.5-Coder-32B-Instruct",  
    messages=[{  
        "role": "user",  
        "content": "Write a Python tutorial on asynchronous web scraping, including code examples and precautions."  
    }],  
    temperature=0.7,  
    max_tokens=4096  
)  

```

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

### Launch Claude Code

Source: https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-ClaudeCode

Starts Claude Code in the terminal after setup is complete. You can then interact with your selected AI model for coding assistance.

```bash
claude

```

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

### JavaScript fetch to get video status

Source: https://docs.siliconflow.com/en/api-reference/videos/get_videos_status

JavaScript example using fetch to POST a requestId to the video status endpoint. Logs the JSON response or error.

```javascript
const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({requestId: '<string>'})
};

fetch('https://api.siliconflow.com/v1/video/status', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

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

### Create virtual environment and install dependencies

Source: https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-DB-GPT

Creates a conda environment with Python 3.10, activates it, and installs DB-GPT with proxy model support. Requires Python >= 3.10.

```bash
# Navigate to the root directory of the DB-GPT source code
cd DB-GPT

# DB-GPT requires Python >= 3.10
conda create -n dbgpt_env python=3.10
conda activate dbgpt_env

# Install dependencies for proxy model support
pip install -e ".[proxy]"

```

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

### Run CodeWhale in terminal

Source: https://docs.siliconflow.com/en/usercases/how-to-use-code-whale-with-silicon-flow-deep-seek-v4-terminal-coding-agent-setup

Launch CodeWhale from any directory or inside a specific project. On first launch, follow the setup steps: press Enter to continue past the welcome screen, choose your preferred language, press 1 to trust and continue, and press Enter to open the workspace.

```bash
#Launch CodeWhale from any directory:
codewhale

#Or run it inside a specific project:
cd path/to/project
codewhale

```

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

### Python requests to get video status

Source: https://docs.siliconflow.com/en/api-reference/videos/get_videos_status

Python example using the requests library to POST a requestId to the video status endpoint. Prints the response text.

```python
import requests

url = "https://api.siliconflow.com/v1/video/status"

payload = { "requestId": "<string>" }
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)
```

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

### TypeSafeClient system_one request example

Source: https://docs.siliconflow.com/en/api-reference/systemone/create-systemone

Creates a TypeSafeClient with a SiliconFlow base URL and API key, then calls system_one with a state string and a map of questions. The example uses Noul, Choice, and Score question types and prints the results.

```python
from typesafe_sdk import Choice, Noul, Score, TypeSafeClient

with TypeSafeClient(
    base_url="https://api.siliconflow.com",
    model="Kev-4B",
) as client:
    result = client.system_one(
        "My withdrawal has failed for three days in a row and support has not replied. Please handle this as soon as possible!",
        {
            "is_urgent": Noul(instructions="Does this message convey urgency?"),
            "department": Choice(
                instructions="Which team should handle this?",
                criteria={
                    "billing": "Invoicing, refunds, and payment issues",
                    "technical": "Bugs, outages, and integrations",
                    "support": "Service attitude and response times",
                },
            ),
            "frustration": Score(
                instructions="How frustrated is the customer?",
                criteria=["Calm", "Slightly annoyed", "Clearly annoyed", "Very angry"],
            ),
        },
    )

print(result.nouls["is_urgent"].noul)
print(result.choices["department"].choice)
print(result.scores["frustration"].score)
```

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

### Successful 200 Response for Get Model List

Source: https://docs.siliconflow.com/en/api-reference/models/get-model-list

Example 200 response for the model list endpoint. Contains an object property set to "list" and a data array with model objects. The shown model has id stabilityai/stable-diffusion-xl-base-1.0, object "model", created 0, and an empty owned_by string.

```json
{
  "object": "list",
  "data": [
    {
      "id": "stabilityai/stable-diffusion-xl-base-1.0",
      "object": "model",
      "created": 0,
      "owned_by": ""
    }
  ]
}
```

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

### Start MindSearch backend with SiliconFlow

Source: https://docs.siliconflow.com/en/usercases/use-siliconcloud-in-mindsearch

Set the SILICON_API_KEY environment variable to the API key obtained earlier, then start the MindSearch backend with the internlm_silicon model format and DuckDuckGoSearch engine.

```bash
# Specify the SiliconFlow API Key
export SILICON_API_KEY=The key copied in the previous steps
# Start
python -m mindsearch.app --lang en --model_format internlm_silicon --search_engine DuckDuckGoSearch
```
Tokens:
Raw