name: flowise-prediction
description: |
  Invoke a Flowise chatflow or agentflow to generate a model response. The
  Prediction endpoint is the primary runtime surface for any Flowise-built AI
  application. Supports streaming, multi-turn history, runtime overrides, RAG
  source documents, tool use, and file uploads.
provider: flowise
api: flowise
authentication:
  type: bearer
  description: HTTP Bearer using a Flowise API key. The chatflow may also have
    its own per-flow API key requirement.
baseUrl: http://localhost:3000/api/v1
operations:
  - id: create-prediction
    method: POST
    path: /prediction/{id}
    summary: Send a message to a flow and receive an AI response
    notes: |
      Body fields include `question`, `streaming`, `overrideConfig`,
      `history`, `uploads`, and `form` (for Agentflow V2). When `streaming`
      is true the response is delivered as Server-Sent Events.
tags:
  - Prediction
  - Inference
  - Runtime
  - Streaming