DeepSeek Models

A drag-and-drop component for integrating DeepSeek's language models. Configure model parameters and connect inputs/outputs to leverage DeepSeek's advanced AI capabilities.

DeepSeek component interface and configuration

API Key Required: A valid DeepSeek API key is required to use this component. You'll need to register for access to DeepSeek's API platform and generate an API key before using this component.

Component Inputs

Input: Text input for the model
Example: "Write a function in Python to implement a binary search algorithm."
System Message: System prompt to guide model behavior
Example: "You are an AI assistant specialized in writing efficient and well-documented code."
Stream: Toggle for streaming responses
Example: true (for real-time token streaming) or false (for complete response)
Model Name: The DeepSeek model to use
Example: "deepseek-chat", "deepseek-coder"
DeepSeek API Key: Your API authentication key
Example: "sk-deepseek-xxxx..."
DeepSeek API Base: API endpoint URL
Example: "https://api.deepseek.com" (Default)
JSON Mode: Toggle for JSON output format
Example: true (for structured JSON responses) or false (for free-form text)

Component Outputs

Text: Generated text output
Example: "```python\ndef binary_search(arr, target):\n left, right = 0, len(arr) - 1\n \n while left = right:\n mid = (left + right) // 2\n \n if arr[mid] == target:\n return mid\n elif arr[mid] target:\n left = mid + 1\n else:\n right = mid - 1\n \n return -1\n```"
Language Model: Model information and metadata
Example: model: deepseek-coder, usage: {prompt_tokens: 45, completion_tokens: 120, total_tokens: 165}

Generation Parameters

Max Tokens

Maximum number of tokens to generate in the response

Default: 4096
Range: 1 to model maximum
Recommendation: Set based on expected response length

Temperature

Controls randomness in the output - higher values increase creativity

Default: 1.0
Range: 0.0 to 2.0
Recommendation: Lower (0.1-0.3) for factual/code responses, Higher (0.7-1.0) for creative tasks

Seed

Random seed for reproducible outputs

Default: 1
Range: Any integer
Recommendation: Set specific values for reproducible results

Model Kwargs

Additional model parameters passed as key-value pairs

Examples:
- top_p: 0.9 (nucleus sampling parameter)
- frequency_penalty: 0.5 (reduces repetition)
- presence_penalty: 0.5 (encourages diversity)
- stop: ["\n\n"] (tokens at which to stop generation)

Available Models

DeepSeek Chat

General-purpose language models for conversational tasks

Models: 
- deepseek-chat (7B parameters)
- deepseek-chat-v2 (latest version)

DeepSeek Coder

Specialized language models for programming tasks

Models: 
- deepseek-coder (6.7B parameters)
- deepseek-coder-instruct (instruction-tuned)

Implementation Example

// Basic configuration
const deepSeekConfig = {
  modelName: "deepseek-chat",
  deepSeekApiKey: process.env.DEEPSEEK_API_KEY,
  systemMessage: "You are a helpful assistant."
};

// Advanced configuration
const advancedDeepSeekConfig = {
  modelName: "deepseek-coder",
  deepSeekApiKey: process.env.DEEPSEEK_API_KEY,
  deepSeekApiBase: "https://api.deepseek.com",
  maxTokens: 2000,
  temperature: 0.3,
  seed: 42,
  jsonMode: true,
  stream: true,
  modelKwargs: {
    top_p: 0.95,
    frequency_penalty: 0.2,
    presence_penalty: 0.2,
    stop: ["\n\n"]
  }
};

// Usage example
async function generateCode(input) {
  const response = await deepSeekComponent.generate({
    input: input,
    systemMessage: "You are an expert programmer. Write clean, well-documented code.",
    modelName: "deepseek-coder",
    temperature: 0.2
  });
  
  return response.text;
}

Use Cases

Code Generation: Create code snippets, functions, and algorithms across multiple programming languages
Technical Documentation: Generate clear explanations and documentation for technical concepts
Conversational AI: Build chat interfaces and virtual assistants
Data Transformation: Leverage JSON mode for structured data formatting
Content Creation: Generate articles, blogs, and other written content

Useful Resources

Best Practices

Use JSON mode when you need structured, parseable outputs
Enable streaming for better user experience with real-time responses
Set a specific seed when you need reproducible results
Adjust temperature based on task type (lower for precision, higher for creativity)
Store API keys securely using environment variables
Monitor your token usage for cost management
Implement proper error handling for API failures
Choose DeepSeek Coder specifically for programming tasks

Documentation