Template Showcase

Multimodal Assistant

Explore the power of multimodal AI. This example shows how to handle images, PDFs, and audio files alongside text prompts using models like GPT-4o.

Clone Repository

ainative.sh/examples/multimodal-assistant

Core Capabilities

Every starter is pre-configured with industry best practices and deep AINative integrations. Clean, modular, and designed to scale from prototype to production.

Vision support for image analysis

Enterprise-ready implementation with full type safety and observability. Optimized for high-throughput streaming and minimal latency.

Document processing (PDF, TXT)

Optimized for Edge runtimes and serverless environments.

Audio transcription and analysis

Secure by default with built-in validation and CSRF protection.

Unified drag-and-drop input

Deep integration with the AINative Reconciler state system for buttery smooth UI updates and zero-flicker streaming.

Technical Stack

Clean, modular, and designed to scale from prototype to production.

Frontend Layer

Reactive hooks and streaming primitives.

TYPESCRIPTREACT 19

Chat.tsx

import { useMultimodal } from "@ainative/react";

export function MultiAssistant() {
  const { uploadFile, submitPrompt } = useMultimodal();

  return (
    <div>
      <Dropzone onDrop={files => files.map(uploadFile)} />
      <Input onSend={submitPrompt} />
    </div>
  );
}

Backend Adapter

Universal protocol handlers for any runtime.

NODE.JSEDGE READY

route.ts

import { MultimodalHandler } from "@ainative/server-node";

export const POST = MultimodalHandler({
  provider: "openai",
  model: "gpt-4o",
  enableVision: true,
  enableAudio: true
});