Skip to main content

​
Overview

This tutorial shows you how to build a Retrieval-Augmented Generation (RAG) chatbot that answers user questions by searching your knowledge base and using the results to generate accurate, context-aware responses.

​
How It Works

​
Step 1: Set Up Your Knowledge Base

First, create a folder for your documentation and ingest your files.

​
Create a Folder

const response = await fetch('https://api.Rayrift.com/v1/folders', {
  method: 'POST',
  headers: {
    'Authorization': 'sk_live_...',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    name: "Product Manuals",
    description: "User guides and documentation"
  })
});

const { id: folderId } = await response.json();
console.log('Folder ID:', folderId);

​
Ingest Documents

Upload your PDF manuals, FAQs, or other documentation: 
// Upload a PDF file
const formData = new FormData();
formData.append('file', pdfFile);
formData.append('payload', JSON.stringify({
  folder_id: folderId
}));

const uploadResponse = await fetch('https://api.Rayrift.com/v1/documents', {
  method: 'POST',
  headers: {
    'Authorization': 'sk_live_...'
  },
  body: formData
});

const document = await uploadResponse.json();
console.log('Document uploaded:', document.id);

​
Step 2: Implement the Search Function

Create a function that searches your knowledge base: 
async function searchKnowledgeBase(query, folderId) {
  const response = await fetch('https://api.Rayrift.com/v1/search', {
    method: 'POST',
    headers: {
      'Authorization': 'sk_live_...',
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      query: query,
      limit: 5,
      folder_id: folderId
    })
  });

  const { results } = await response.json();
  return results;
}

​
Step 3: Build the Chatbot Logic

Combine search results with an LLM to generate answers: 
async function answerQuestion(userQuestion, folderId) {
  // Step 1: Search the knowledge base
  const searchResults = await searchKnowledgeBase(userQuestion, folderId);
  
  // Step 2: Build context from search results
  const context = searchResults
    .map((hit, index) => `[${index + 1}] ${hit.text}`)
    .join('\n\n');
  
  // Step 3: Create prompt for LLM
  const prompt = `You are a helpful assistant. Answer the user's question using only the information provided in the context below. If the context doesn't contain enough information, say so.

Context:
${context}

User Question: ${userQuestion}

Answer:`;

  // Step 4: Call your LLM (OpenAI, Anthropic, etc.)
  const llmResponse = await callLLM(prompt);
  
  return {
    answer: llmResponse,
    sources: searchResults.map(hit => ({
      text: hit.text.substring(0, 100) + '...',
      score: hit.score
    }))
  };
}

// Example LLM integration (using OpenAI)
async function callLLM(prompt) {
  const response = await fetch('https://api.openai.com/v1/chat/completions', {
    method: 'POST',
    headers: {
      'Authorization': `Bearer ${process.env.OPENAI_API_KEY}`,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      model: 'gpt-4',
      messages: [{ role: 'user', content: prompt }],
      temperature: 0.7
    })
  });

  const data = await response.json();
  return data.choices[0].message.content;
}

​
Step 4: Create the Chat Interface

Build a simple chat interface: 
// Express.js example
app.post('/api/chat', async (req, res) => {
  try {
    const { message, folderId } = req.body;
    
    const result = await answerQuestion(message, folderId);
    
    res.json({
      answer: result.answer,
      sources: result.sources
    });
  } catch (error) {
    res.status(500).json({ error: error.message });
  }
});

​
Complete Example

Here’s a complete Node.js/Express implementation: 
const express = require('express');
const fetch = require('node-fetch');
const app = express();

app.use(express.json());

const Rayrift_API_KEY = process.env.Rayrift_API_KEY;
const Rayrift_API_URL = 'https://api.Rayrift.com/v1';
const OPENAI_API_KEY = process.env.OPENAI_API_KEY;

async function searchRayrift(query, folderId) {
  const response = await fetch(`${Rayrift_API_URL}/search`, {
    method: 'POST',
    headers: {
      'Authorization': Rayrift_API_KEY,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      query,
      limit: 5,
      folder_id: folderId
    })
  });
  
  const data = await response.json();
  return data.results || [];
}

async function generateAnswer(question, context) {
  const prompt = `Answer the following question using the provided context. If the context doesn't contain enough information, say so.

Context:
${context}

Question: ${question}

Answer:`;

  const response = await fetch('https://api.openai.com/v1/chat/completions', {
    method: 'POST',
    headers: {
      'Authorization': `Bearer ${OPENAI_API_KEY}`,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      model: 'gpt-4',
      messages: [{ role: 'user', content: prompt }]
    })
  });

  const data = await response.json();
  return data.choices[0].message.content;
}

app.post('/chat', async (req, res) => {
  const { message, folderId } = req.body;
  
  try {
    // Search knowledge base
    const results = await searchRayrift(message, folderId);
    
    // Build context
    const context = results.map(r => r.text).join('\n\n');
    
    // Generate answer
    const answer = await generateAnswer(message, context);
    
    res.json({
      answer,
      sources: results.map(r => ({
        snippet: r.text.substring(0, 150),
        score: r.score
      }))
    });
  } catch (error) {
    res.status(500).json({ error: error.message });
  }
});

app.listen(3000, () => {
  console.log('Chatbot server running on port 3000');
});

​
Best Practices

Improve Search Quality: Use specific folder IDs to scope searches to relevant documentation.
Handle Edge Cases: Check if search results are empty and provide appropriate fallback responses.
Show Sources: Display source snippets to users so they can verify information.
Rate Limiting: Implement rate limiting on your chat endpoint to prevent abuse.

​
Next Steps