Upstage Document ChatInstall

Summary

Send a question or instruction to chat about one or more uploaded documents using the selected Upstage Solar model and receive a text response answering or addressing the query.

Inputs

• Model — Select the language model to use for the chat, either 'Genius' (balanced performance) or 'Turbo' (optimized for speed). Defaults to 'Genius'.
• File ID(s) (required) — Comma-separated list of file IDs corresponding to previously uploaded documents to be used in the chat.
• Query (required) — The question or instruction you want to ask about the provided documents.
• Options — Additional configuration for the response, including:
• Conversation ID: to continue an existing conversation by ID.
• Stream: whether the response should be streamed (boolean).
• Reasoning: settings for the response reasoning effort level (low, medium, high) and whether to include a reasoning summary (auto, enabled, disabled).

Output shape

a single JSON object per input item containing the response text, conversation ID if provided by the API, the original query, and the full API response object

Returns the first message text from the API response content under 'content', the conversation ID if continuing a conversation, and includes the complete raw response under 'full_response'. Errors are returned per item with error details if 'Continue On Fail' is enabled.

Examples

Example 1: Ask a question about specific uploaded documents using the default 'Genius' model.

Provide one or more File IDs, enter the query text, optionally configure conversation continuity or reasoning details, then execute to receive the chat response text and conversation metadata.