Unified AI Python LLM Connector2025-12-19T09:39:57+00:00
One connector to rule them all

Unified AI Python LLM Connector

Licence Apache 2.0 • Commercial Use Allowed • 100% Free

The Universal Python LLM Connector for ChatGPT, Ollama, Qwen, DeepSeek, Gemini & many more. Inject ANY model at runtime and build Multimodal Agents in minutes. 100% Free & Open Source.

Ideal-ai universal llm connector

One connector to rule them all 🧙‍♂️

One interface for OpenAI ChatGPT, Google Gemini, Anthropic Claude, DeepSeek, Ollama, Alibaba Qwen, Infomaniak, Moonshot, Mistral, Hugging Face, Perplexity, Apertus and more. Inject ANY custom model at runtime. Switch providers instantly without rewriting your application code.

Launch the FULL Multi-Modal Demo on Colab :


Launch the FULL Multi-Modal Unified AI LLM Connector Demo on Colab!

Built for the Agentic Era. Native support for Smolagents & LangChain. Seamlessly handle Wan 2.1 (Video), Whisper (Audio) & Flux with automatic async polling.

Get started in seconds. Inject custom models at runtime without rewriting your code. Features smart configuration and robust error handling for production.

pip install ideal-ai

The Universal Multi-LLM AI Connector

Open Source, 100% free, Plug & Play, Local-first.
Stop wrapping, start building. ❤️
pip install ideal-ai

🤗
Native Smolagents & LangChain Ready

Empower your agents with any brain. Don’t lock your autonomous agents to a single vendor. Whether you use the cutting-edge Smolagents (native wrapper) or the industry-standard LangChain (direct workflow integration), plug any model from local Ollama to cloud-based GPT-5 instantly into your agent workflows.

Empower your agents with any brain. Don’t lock your autonomous agents to a single vendor. Whether you use the cutting-edge Smolagents (native wrapper) or the industry-standard LangChain (direct workflow integration), plug any model—from local Ollama to cloud-based GPT-4—instantly into your agent workflows.

🧙‍♂️
Switch models, not code

Scale instantly from local prototyping (Ollama) to SOTA production (OpenAI, Mistral). The connector lets you dynamically inject any LLM into your existing AI agents.
Zero refactoring, total flexibility.


Master every modality

Text is just the beginning. Orchestrate Vision, Speech, Image, and Video generation with zero friction. One unified API to rule all senses. Stop wrestling with fragmented SDKs focus on building immersive, multi-modal experiences.

🔭
Explore Without Boundaries

Great ecosystems shouldn’t feel like silos. Moving between providers usually takes energy and time. We bridge the gap. Test, iterate, and innovate across all environments.
Open your project to infinite possibilities.

Great ecosystems shouldn’t feel like silos. Moving between providers usually takes energy and time. We bridge the gap. Test, iterate, and innovate across all environments.
Open your project to infinite possibilities.

Key Features

  • ✨ Unified interface for 15+ AI providers

    One API to rule them all: OpenAI, DeepSeek, Ollama, Google, Anthropic, Alibaba, and more.

  • 🔌 Multi-modal: Text, Vision, Audio, Image, Video

    Handle everything in one place: Text, Vision, Audio (STT), Image Gen, Video Gen (Wan 2.1) & Speech (TTS).

  • 🚀 Dynamic Model Injection (Cloud vs Local)

    Switch between Cloud & Local models instantly without rewriting a single line of code.

  • 🤖 Agent & Workflow Ready

    Native wrapper for Smolagents and fully compatible with LangChain & LangGraph.

  • 🛡️ Production-Grade Reliability

    Robust error handling with automatic async polling for long tasks (Video/Whisper).

  • 🎙️ Native Voice Chat Pipeline

    Ready-to-use module for full audio-to-audio interaction (Talk with your AI).

  • 💼 100% Open Source & Free

    Apache 2.0 License. No hidden fees. Commercial use allowed.

  • 📦 Standard Python Package

    Simple installation via pip install ideal-ai. No complex setup.

FAQ LLM AI Connector

Can I use this LLM Connector for commercial projects ?2025-12-11T10:25:21+00:00

Yes. The project is licensed under Apache 2.0, which allows you to use it in commercial, closed-source, and enterprise applications without restriction. It is 100% free, even for business use.

Who is this Connector built for ?2025-12-11T10:23:19+00:00

It is designed for Python Developers and AI Engineers who need clean, production-ready architecture.

It is also perfect for Students & AI Beginners who want to accelerate their learning: experiment with different models (switch from OpenAI to Local Llama 3) instantly, focusing on the concepts rather than struggling with complex API documentation.

Why use Ideal-AI instead of official SDKs (OpenAI/Google) ?2025-12-11T10:21:01+00:00

To avoid Vendor Lock-in. If you use the official OpenAI SDK, you are trapped. If they raise prices or change policies, rewriting your code takes days. With Ideal-AI, switching to a cheaper or better competitor (like DeepSeek) takes 1 second.

Can I prototype locally for free before going to the Cloud ?2025-12-11T10:19:35+00:00

Absolutely! This is the best use case. Build and test your agent locally with Ollama (free, $0 cost). Once it works perfectly, switch to GPT-4o or Claude 3.5 for production just by changing the model_id.

No code rewrite needed.

Is Ideal-AI really free to use ?2025-12-11T10:25:44+00:00

Yes, the library itself is 100% Free and Open Source (Apache 2.0 License). You don’t pay us anything.

Note: If you use paid cloud providers like OpenAI or Anthropic, you still pay them directly. However, using local models (like Ollama) is completely free.

How do I install it ?2025-12-06T16:09:00+00:00

pip install ideal-ai

Is my data stored locally ?2025-12-06T16:07:47+00:00

Yes; with Ollama local-first architecture.

Can I use the Connector with Smolagents ?2025-12-06T16:06:57+00:00

Yes; built-in wrapper enables agent workflows immediately.

Does the Connector support local models ?2025-12-06T16:06:13+00:00

Yes; full Ollama integration for on-premise deployments.

Do I need to rewrite my code for each provider ?2025-12-11T10:26:09+00:00

No; switch providers instantly with a few lines of code.

What is Universal AI LLM Connector ?2025-12-06T16:03:46+00:00

A Python connector unifying 15+ AI providers under one API.

The Wall of Giants

Standing on the shoulders of giants

“Built by Gilles Blanchet, founder of Idealcom.ch and IA-Agence.ai. Backed by 18+ years of digital expertise, this connector is designed to simplify the AI landscape for developers and future-proof your unified infrastructure.”

Ready to Unify Your AI Stack?

pip install ideal-ai

“The production-ready universal connector your infrastructure deserves. Open source, fully extensible, and built for developers.”

One Connector to Rule Them All.

Team IA