Skip to main content

Quick start

This guide helps you integrate intellinode in your Node.js project for basic tasks like chatbots, one-line generation and in-memory semantic search.

Installation​

Install IntelliNode using npm:

npm install intellinode

Chatbot​

Easily integrate and switch between chatbot providers:

const { Chatbot, ChatGPTInput, AnthropicInput, GeminiInput, SupportedChatModels } = require('intellinode');

async function callChatbot(provider, apiKey, input) {
  input.addUserMessage('What is the capital of France?');

  // create chatbot instance
  const chatbot = new Chatbot(apiKey, provider);
  const responses = await chatbot.chat(input);

  console.log(responses[0]);
}

// call GPT-5.5 (the default openai model)
callChatbot(SupportedChatModels.OPENAI, 'your_openai_api_key', new ChatGPTInput('You are a helpful assistant.'));

// call Claude Sonnet 5 (the default anthropic model)
callChatbot(SupportedChatModels.ANTHROPIC, 'your_anthropic_api_key', new AnthropicInput('You are a helpful assistant.'));

// call Gemini with a specific model
callChatbot(SupportedChatModels.GEMINI, 'your_gemini_api_key', new GeminiInput('You are a helpful assistant.', { model: 'gemini-3.6-flash' }));

Replace your_<model>_api_key with your actual API key and explore other supported models. Every input class accepts { model, temperature, maxTokens } in its second argument.

The same chatbot streams the answer, runs your tools and returns JSON that matches a schema:

// stream
for await (const chunk of chatbot.stream(input)) process.stdout.write(chunk);

// JSON that matches a schema
const input = new ChatGPTInput('Answer as JSON.', { responseSchema: { type: 'object', properties: { city: { type: 'string' } } } });
input.addUserMessage('Where is the Eiffel Tower?');
const data = await chatbot.chatJson(input);   // { city: 'Paris' }

See Tool calling and Structured output for the details.

One-line generation​

The Gen functions do a complete task in one call, with any provider:

const { Gen } = require('intellinode');

const code = await Gen.generate_component('a pricing card with a CTA button', openaiKey, 'openai', { framework: 'react', styling: 'tailwind' });
const sql = await Gen.generate_sql('top 10 customers by order total', anthropicKey, 'anthropic', { dialect: 'postgresql' });
const review = await Gen.review_code(code, openaiKey);   // { summary, score, issues }

The full list is in the Gen section.

Find relevant items in your data using semantic similarity:

const { SemanticSearch } = require('intellinode');

const apiKey = 'your_api_key';
const search = new SemanticSearch(apiKey);

// prepare the input
const pivotItem = 'Example search item';
const searchArray = ['Item 1', 'Item 2', 'Item 3', 'Item 4'];
const numberOfMatches = 3;

// perform the search
const results = await search.getTopMatches(pivotItem, searchArray, numberOfMatches);