Speaker 1: Hi, my name is Jose Francisco, and today I'll be showing you how to accurately transcribe any audio you want as quickly and efficiently as possible. In other words, this is a tutorial on how to use DeepGram. This tutorial is meant to be extremely quick, hence the background music. In fact, the piece playing right now is Beethoven's Moonlight Sonata, 3rd movement. According to various YouTube videos, this piece is around 6-7 minutes long, but with this quick start tutorial, we'll go over how to transcribe any audio you want before the music ends. All you have to do is follow our notebook. Link in the description. Ready? Let's go. First things first, open up the notebook. Now, make a copy of the notebook, like this. This tutorial will assume that you're using Google Colab, but even if you're using Jupyter Notebooks or running this on VS Code, the general instructions should be about the same. Alright, now that you've made a copy of the notebook, let's run the first cell. To run a cell, simply click the play button, or press shift and enter at the same time. Or shift and return if you're on a Mac. This cell simply installs dependencies using pip. Oh yeah, we're working in Python here. Give it a few moments, and you'll see some colorful text, like this. Now, for some people out there, you may need to use pip3 instead of pip depending on your setup, but the output should remain the same. Alright, next, grab any audio file you want. You can even grab multiple. But for the sake of this tutorial, I'm going to use the first chapter of the audiobook from Emma by Jane Austen. I have the mp3 already downloaded onto my laptop, so I'll just upload that into Google Colab, like this. And if you're using Jupyter or VS Code, use the Analysis Upload feature of that platform. Now this audio is ready to go. The next cell to run is the transcription cell. All we have to do here is fill in some variables. First I'll fill in my API key, which is blurred and blacked out here for security reasons. But you can get your own API key by signing up for DeepGram and creating an API key with your own account. Again, link is in the description. Now the next variable you have to modify is the MIME type. Right now the MIME type is set to mp3. Luckily for me, my audio is already in mp3, so I don't have to change anything. But you can totally change it if you're working with an mp4, m4a, or anything else. The most common audio formats and encodings that we support are featured on screen right now. Now I modify the directory variable. This variable should be set to the name of the folder that contains the audios that we wish to transcribe. Since I just placed my audio in the current directory, the classic dot should suffice. But if you, for example, put your audios inside the default sample data folder, or if you went the extra mile and created your own folder to host your audios, then set this variable to the name of that folder. And we should be good to go. Run this cell and you should see a JSON appear in the folder on your left. Note, there may be a slight delay between when your cell finisher is running and when the JSON appears in the folder, but that's just a little google colab quirk. Rest assured, your output JSON file should appear in less than a minute. Alright, and with the JSON ready, we can take a peek at it. Whoa, pretty colors. A quick glance at the JSON reveals that we provide metadata, word level timestamp, and confidence levels for every single word. But if you just want the transcript, fret not. We provide the entire transcript as one of the fields in the JSON. It's just a string. And running the final cell will allow us to see this transcript. I wrote this function to make each printed line a sentence. Boom. You have a printed transcript right there. Let's find a fun little sentence and read it.

Speaker 2: She was the youngest of the two daughters of a most affectionate, indulgent father, and had, in consequence of her sister's marriage, been mistress of his house from a very early period. Beautiful.

Speaker 1: It practically matches the actual text, give or take a few comments. And that's how you use DeepGram as quickly as possible. Feel free to rewrite the code in the notebook as much as you desire. Or if you just want to write code with DeepGram yourself, check out the DeepGram SDK, or Software Development Kit. We have SDKs for Node, Python, Go, and much more. Not to mention many other features. For example, a quick look at our documentation reveals that you can also transcribe audio straight from a URL. Not only that, but we offer live transcription as well. And with the same SDK, you can manage your projects, diarize your transcript so you can label every speaker, and even use models that are optimized for other types of audios. Audios like phone calls, meetings, voicemails, and even conversations with AI. Just look at some of the cool demos our users have come up with. You can code a website with your voice, drive a car with your voice, create a live subtitles badge, and much, much more. It all starts with this notebook. But that's DeepGram in a nutshell. A quick, easy-to-use API with intuitive documentation written by humans for humans. Alright, what's my time? Nailed it.