DeerAPI Plus icon

DeerAPI Plus

AI 全能工具箱：对话、思维链、图片、视频、向量、提示词优化、虚拟试穿 | AI Toolkit: Chat, Thinking, Image, Video, Embeddings, Prompt, Virtual Try-On

Actions11

💬 对话 / Chat → 生成 / Generate

AI-generated

Summary

Send a user prompt to a DeerAPI chat model and retrieve the model's response.

Inputs

  • mode — Choose a preset mode or custom model; defaults to 'recommended' if omitted. When set to 'custom', the 'model' and 'customModel' parameters are used to specify the model ID directly.
  • model — The model to use for the chat. Loaded dynamically from DeerAPI when the mode is not 'custom'; defaults to 'gemini-2.5-flash'. When mode is 'custom', this parameter selects the model; if set to '__custom', the 'customModel' parameter provides the ID.
  • customModel — Enter a custom model ID. Only shown when mode is 'custom' and model is '__custom'.
  • userPrompt (required) — The message to send to the model. This is the user's input prompt.
  • additionalOptions — Optional collection of additional settings including systemPrompt (default: 'You are a helpful assistant.'), temperature (0-2, default 0.7), maxTokens (1-16384, default 2048), binaryPropertyName (comma-separated property names for multimodal images), simplify (boolean, default true), and extraBodyFields (JSON string merged into the request body).

Output shape

The node returns a single item with a 'json' field containing the operation result.

The 'simplify' parameter (default true) controls whether the output includes only key fields or the full response. If binaryPropertyName is provided and binary image data exists in the input, it is embedded as multimodal content in the user prompt before sending to the model. The resolved model ID is always included in the return JSON.

Examples

Example 1: Basic chat generation with default mode.

mode: 'recommended' (default), userPrompt: 'Hello, how are you?'. The model resolves to 'gemini-2.5-flash' and the node returns a simplified JSON with success, operation, model, content, finish_reason, and processing_time_ms.

Example 2: Chat generation with custom model and image input.

mode: 'custom', model: '__custom', customModel: 'gpt-4o', userPrompt: 'Describe this image:', binaryPropertyName: 'data'. The node extracts binary image data from input, embeds it as image_url blocks, and sends the multimodal prompt. The return includes the model response content and processing time.

Links

Discussion