Chat UI

The One Key document service is no longer available. To answer from your own documents in code, use the Assistant. As a UI alternative, IntelliChat lets you connect any supported model with your own key, and chat with it from the browser without writing code. It is open source and built on the same Chatbot you use in code.

Connect a Model

1. Open chat.intellinode.ai and open the settings.
2. Choose the Chat, Images, Voice or Code tab.
3. Select the provider and the model.
4. Paste your API key and save.

Each tab keeps its own key, so you can chat with one provider and generate images or voice with another.

Supported Providers

Group	Providers
Cloud	OpenAI, Anthropic, Google Gemini, Cohere, Mistral, Replicate (Llama), Azure OpenAI
OpenAI-compatible	OpenRouter, Groq, DeepSeek
Local and self-hosted	Ollama, LM Studio, vLLM

Features

• Stream replies and stop them at any time.
• Generate images with the image button or /image.
• Attach an image and ask about it.
• Dictate messages with the microphone and listen to the replies.
• Paste a GitHub link to ask about a repo, file, issue or pull request.
• Connect a GitHub repo, or a local folder when running locally, and the assistant reads the code and shows each step.

Run It Locally

Run IntelliChat on your machine to use local models such as Ollama, LM Studio or a vLLM server, or to load your keys from a .env file:

git clone https://github.com/intelligentnode/IntelliChat.git
cd IntelliChat/intellichat
npm install
npm run dev

Open http://localhost:3000. Optionally copy .env.example to .env and add your keys, or add them from the settings.

The Same Chat in Code

The UI settings map to the Chatbot arguments, so you can move from the UI to your app with the same provider and model:

const { Chatbot, ChatGPTInput } = require("intellinode");

const bot = new Chatbot(openaiKey, "openai");

const input = new ChatGPTInput("You are a helpful assistant.", { model: "gpt-5.5" });
input.addUserMessage("List the included features in the vector database contract.");

const responses = await bot.chat(input);
responses.forEach(response => console.log("- " + response));

See Get started for every provider, and OpenAI-compatible for OpenRouter, Groq, DeepSeek and local servers.