DeepSeek Block

The DeepSeek Block integrates DeepSeek's efficient and high-performing language models into your flow. Similar to other AI blocks, it provides seamless history tracking and parameter customization.

Configuration

Field	Required	Description
API Key	✅	Your DeepSeek API key.
Model	✅	The DeepSeek model to invoke (e.g., `deepseek-chat`, `deepseek-coder`).
User Prompt	✅	The text query for the model.
System Prompt	—	System instructions to guide the model's output.
Enable History	—	Enable to string together multi-turn interactions.

History Settings

If Enable History is active:

Storage Object Variable (Required): The object variable storing the conversation data.
Chat ID: Used to separate different user sessions.
Message Count: Defines how much context is sent.

Advanced Settings

Toggle Advanced Settings for model tuning:

Temperature (0 to 2): Adjusts creativity.
Max Output Tokens: Defines maximum response length.
Top P: Used for nucleus sampling.
Frequency & Presence Penalty: Penalties to mitigate repetitive outputs.

Output Variable

Specify a Variable Name to store the result.

Success Payload

{
  "success": true,
  "chat_id": "chat_1234abc",
  "message": "The assistant's response...",
  "reasoning_content": "DeepSeek thought process...",
  "raw_response": {
    "id": "chatcmpl-123",
    "object": "chat.completion",
    "choices": [...]
  }
}

Error Payload

If an error occurs, the variable will be populated with:

{
  "error_key": "invalid_api_key",
  "error_message": "The provided DeepSeek API Key is invalid or expired."
}

Error Handling

API_KEY_MISSING / SECRET_ERROR — Missing or invalid API key.
MODEL_MISSING — No model was specified.
PROMPT_MISSING — The user prompt was empty.
HISTORY_VAR_MISSING — History was enabled but no storage object was provided.
RUNTIME_ERROR — An unexpected issue occurred during the API call.
invalid_api_key / insufficient_quota / rate_limit_exceeded / context_length_exceeded / model_not_found — Standard DeepSeek API errors.

DeepSeek

On this page