Introducing OmiAI: The Future of Intelligent AI Model Selection

30 Jan 2025 • 3 min read

Angel

Share this article

Introducing OmiAI: The Future of Intelligent AI Model Selection

In the early days of LLMs, things were simpler. GPT-3.5 reigned supreme, and developers only had to integrate one model, one framework, and one standard. Fast forward to today, and the landscape has transformed dramatically. With a surge of new models like DeepSeek R1, Gemini Flash 8B, and Sonnet 3.5—each excelling in different areas—developers now face the challenge of choosing the right model for the right task.

Introducing OmiAI, an opinionated, highly optimized LLM SDK designed to eliminate the complexity of model selection. Built on Vercel’s AI SDK and powered by JigsawStack’s embedding model and AI tools, OmiAI automatically selects the best AI model based on the task at hand, seamlessly integrating reasoning, tool-calling, and multi-modal support to deliver the most accurate and efficient results.

How OmiAI Works

Instead of manually selecting an AI model or managing schemas, OmiAI simplifies the process by dynamically choosing the best option. Here’s a breakdown of its execution flow:

Automatic Model Selection – Based on the user’s prompt, OmiAI picks the most suitable model from a pre-selected list.
Intelligent Tool-Calling – If additional context or actions are needed (e.g., web search, OCR, AI scraping), OmiAI automatically invokes relevant tools.
Reasoning Layer Integration – For complex queries, OmiAI leverages DeepSeek R1 to enhance decision-making and response accuracy.
Multi-Modal Support – OmiAI can process PDFs, images, audio, CSVs, JSON, and more, allowing for comprehensive AI-driven solutions.
Final Execution & Response Optimization – All intermediate outputs are processed through the selected LLM, ensuring a seamless and accurate response.

Why OmiAI?

Effortless Optimization

Unlike traditional AI frameworks that require developers to manually switch between models, OmiAI removes the guesswork by selecting the best model every time. Whether it’s a coding-heavy task, a long-form content generation request, or a data extraction problem, OmiAI ensures you’re always using the most effective model.

Built-In Reasoning & Tool-Calling

Many LLMs struggle with complex problem-solving. OmiAI bridges this gap by integrating reasoning models like DeepSeek R1 and dynamic web search to enrich responses with real-world, up-to-date information.

Seamless Multi-Modal Processing

OmiAI is not just for text—it’s fully capable of handling files, images, and structured data. Whether you need AI-powered OCR for invoices, summarization for PDFs, or embedding for text and documents, OmiAI makes it effortless.

Getting Started with OmiAI

With a single npm install, developers can integrate OmiAI into their projects without worrying about complex configurations. It includes rate-limiting fallbacks, retries, and structured output support (via Zod) to ensure stability and reliability.

Installation

npm i omiai

Basic Usage

import { createOmiAI } from "omiai";

const omi = createOmiAI();
const result = await omi.generate({ 
    prompt: "What is the meaning of life?" });
console.log(result?.text);

Structured Output with Zod

import { z } from "zod";
const result = await omi.generate({
  prompt: "How many r's are there in the word 'strawberries'?",
  schema: z.object({ 
      answer: z.number().describe("The answer to the question") }),
});
console.log(result?.object);

Multi-Modal Support

const result = await omi.generate({
  prompt: [
    {
      role: "user",
      content: [
        { type: "text", 
            data: "Extract the total price from the image" },
        { type: "image", 
            data: "https://example.com/image.jpg", mimeType: "image/jpg" },
      ],
    },
  ],
  schema: z.object({ 
    total_price: z.number().describe("Total price of items") }),
});
console.log(result?.object);

What’s Next for OmiAI?

OmiAI is built to stay opinionated—meaning you don’t pick the model; it picks the best one for you. The focus remains on efficiency, minimal configurations, and an optimized developer experience.

While the project is still in its early stages, expect breaking changes, ongoing improvements, and contributions from the community to refine its capabilities.

We want your feedback! If you have suggestions on how OmiAI can be improved or simplified, feel free to open a discussion or PR on GitHub.

🔗 Check out the repo: OmiAI on GitHub

OmiAI isn’t just another AI SDK—it’s a smarter way to build with AI. Try it today and experience automated AI model selection at its finest!

👥 Join the JigsawStack Community

Have questions or want to show off what you’ve built? Join the JigsawStack developer community on Discord and X/Twitter. Let’s build something amazing together!

Share this article