Chat → Complete
AI-generatedOverview
This node interacts with SiliconFlow AI models to generate chat completions. It is designed to create conversational AI responses based on input messages or a simple prompt. This is useful for building chatbots, virtual assistants, or any application requiring AI-generated conversational text. Users can specify the model, input messages or prompt, and various parameters to control the response generation.
Use Case Examples
- Generating a customer support chatbot response based on user messages.
- Creating an AI assistant that responds to user queries with detailed or simple text outputs.
- Using different models to experiment with response styles and capabilities.
Properties
| Name | Meaning |
|---|---|
| Model | The AI model used to generate the chat completion. Supports models that can handle tool calls. |
| Messages | A collection of messages forming the conversation history. Each message has a role (system, user, assistant) and content. Used as input for the chat completion. |
| Prompt | A simple prompt text alternative to providing multiple messages. Used as a single user message if messages are not provided. |
| Output Mode | Determines the format of the output. 'Simple' returns only the message content as a string, while 'Detailed' returns a structured object including message content, usage statistics, and metadata. |
| Additional Fields | Optional parameters to customize the chat completion generation, including max tokens, temperature, top_p, top_k, min_p, frequency penalty, number of generations, thinking mode, thinking budget, stop sequences, streaming option, and response format. |
Output
JSON
message- The generated message content from the AI model.model- The model identifier used for generating the completion.finishReason- Reason why the generation finished (e.g., stop sequence, max tokens).usage- Token usage statistics for the request.reasoning- Optional reasoning content if provided by the model.toolCalls- Optional tool call metadata if the model used tools._rawResponse- The full raw response from the API for debugging or extended use.
Dependencies
- Requires an API key credential for SiliconFlow API access.
- Uses HTTP POST requests to the SiliconFlow API endpoint for chat completions.
Troubleshooting
- Error if neither messages nor prompt is provided: Ensure at least one of these inputs is set.
- API errors may occur if the model ID is invalid or the API key is incorrect.
- Streaming option requires proper handling of Server-Sent Events in the client.
- Stop sequences must be comma-separated strings; invalid formats may cause errors.