Step by Step Guide to Building a Fireflies.ai API Integration in JS

Aug 14, 2024 • 6 minute read

Introduction

Hey there, fellow developer! Ready to supercharge your app with some AI-powered transcription and conversation analysis? Let's dive into integrating the Fireflies.ai API into your JavaScript project. This nifty tool can transcribe your audio, analyze conversations, and even let you search through all that juicy data. Exciting, right?

Prerequisites

Before we jump in, make sure you've got:

Node.js and npm (you're probably nodding already)
A Fireflies.ai API key (if you don't have one, go grab it!)
Axios or your favorite HTTP client library

Setting up the project

Let's get this show on the road:

mkdir fireflies-integration
cd fireflies-integration
npm init -y
npm install axios dotenv

Authentication

First things first, let's keep that API key safe:

// .env
FIREFLIES_API_KEY=your_api_key_here

// index.js
require('dotenv').config();
const axios = require('axios');

const api = axios.create({
  baseURL: 'https://api.fireflies.ai/graphql',
  headers: {
    'Authorization': process.env.FIREFLIES_API_KEY,
    'Content-Type': 'application/json'
  }
});

Making API requests

Now, let's talk to Fireflies:

async function makeRequest(query, variables = {}) {
  try {
    const response = await api.post('', { query, variables });
    return response.data;
  } catch (error) {
    console.error('API request failed:', error);
    throw error;
  }
}

Implementing key Fireflies.ai API features

Transcription

Upload that audio and get it transcribed:

async function uploadAudio(fileUrl) {
  const query = `
    mutation($input: UploadAudioInput!) {
      uploadAudio(input: $input) {
        id
        status
      }
    }
  `;
  const variables = { input: { url: fileUrl } };
  return makeRequest(query, variables);
}

async function getTranscription(transcriptId) {
  const query = `
    query($id: ID!) {
      transcript(id: $id) {
        text
        speakers {
          name
          words
        }
      }
    }
  `;
  const variables = { id: transcriptId };
  return makeRequest(query, variables);
}

Conversation analysis

Let's extract some insights:

async function analyzeConversation(transcriptId) {
  const query = `
    query($id: ID!) {
      transcript(id: $id) {
        summary
        actionItems
        topics
      }
    }
  `;
  const variables = { id: transcriptId };
  return makeRequest(query, variables);
}

Search functionality

Find what you need in a snap:

async function searchTranscripts(keyword) {
  const query = `
    query($keyword: String!) {
      search(query: $keyword) {
        transcripts {
          id
          title
          snippets
        }
      }
    }
  `;
  const variables = { keyword };
  return makeRequest(query, variables);
}

Error handling and best practices

Always be prepared:

function handleApiError(error) {
  if (error.response) {
    console.error('API Error:', error.response.data);
    // Handle rate limiting
    if (error.response.status === 429) {
      // Implement exponential backoff
    }
  } else {
    console.error('Network Error:', error.message);
  }
}

Testing the integration

Don't forget to test! Here's a quick example using Jest:

test('uploadAudio should return a transcript ID', async () => {
  const result = await uploadAudio('https://example.com/audio.mp3');
  expect(result.data.uploadAudio.id).toBeDefined();
});

Optimizing performance

Keep things speedy:

const NodeCache = require('node-cache');
const cache = new NodeCache({ stdTTL: 600 });

async function getCachedTranscription(transcriptId) {
  const cacheKey = `transcript_${transcriptId}`;
  let transcript = cache.get(cacheKey);
  if (!transcript) {
    transcript = await getTranscription(transcriptId);
    cache.set(cacheKey, transcript);
  }
  return transcript;
}

Conclusion

And there you have it! You've just built a solid Fireflies.ai API integration. With this foundation, you can transcribe audio, analyze conversations, and search through your data like a pro. The possibilities are endless – maybe build a meeting summary bot or a conversation insights dashboard?

Additional resources

Now go forth and build something awesome! Remember, the sky's the limit when you've got AI on your side. Happy coding!