12 Days of Tella - Day 8: API & MCP 🤖

Today is day eight, and today I have a special guest.

Tella co founder and CTO Michiel.

Hello.

Michiel's here because, I probably won't do the best job of explaining what, what we are launching, today.

Michiel.

What are we launching?

we're going to release our own public API.

Nice.

This is really cool because it's the first time you and, can extend Tella, to fit into your own workflows.

You can access videos, playlists, you can grab transcripts, you can get the chapters of your video, you can access thumbnails and a bunch more.

So if you head to docs.tella.com you can see all of our documentation, how to get started, all the different parts of it.

And if you head to your settings in your account, you can go and generate your API key there.

You can kick that process off here.

We also got web hooks.

Michiel, Michiel, what's a webhook?

it basically means that we will make an API call to your API endpoint when certain events happen, like video created, transcription ready, export ready, that sort of stuff.

some of our customers have already started building what with the API.

So we'll, hand it over to Connor, who's got some, cool workflows set up already.

Hey, what's going on?

So, absolutely loving the Tella API.

here are two ways that I have been using it over the last few days.

the first thing that I did was I created a database for all of my Tella videos.

So you can see in here, I've imported it all.

I got access to my thumbnails, my embed URLs, the chapters, the transcript, all that type of stuff.

And now that I've got it in here, I can actually use this in a lot of different ways.

First of all, it's a lot easier for me to find certain types of content that I might want to repurpose.

But also I can now also update information inside of Tella directly from here.

So if I for example wanted to change say the name of a video title, instead of having to click around in Tella, I can actually just come in here and I can go ahead update it.

Another thing that I can do in here that's really exciting is I can add it to different playlists.

And I've also set up an automation that basically reviews all of my different videos that are not part of playlists and, and basically add it to a playlist if I already have a playlist that can be used.

one of the things that I have been so excited to try to see how easy it is, is to see if I can create an end to end, workflow for taking a Tella video and turning it into a blog post.

And let me tell you, it works absolutely beautifully.

Now to walk you through how it works, I'll go ahead, I'll film a video like this and then as soon as I'm done with the video, what I'll do is I'll just go ahead and add it to a playlist.

Now I created a playlist called Connor's Blog.

So jump in here and I'll click on Add Playlist Connor's Blog.

Now in the background I've set up a workflow inside of make that is monitoring when I add videos to any of my playlists.

And so in this case I've got a webhook that is monitoring whether or not I've added a video, to the Connors Blog playlist.

And as soon as a new video is added there, it will go through the step by step process of first of all, retrieving the information from Tella about the video that's been added.

So in this case you can see I created a video, gives me the name, gives me the description that's generated.

It gives me access to all the thumbnails, it gives me access to the transcript, basically all the things that you could get in Tella.

But instead of having to copy and paste things over, I can actually just do it automatically.

And then as soon as that is done, I actually plug it into an OpenAI prompt that takes all of the information, and turns it into a blog post.

So I just feed in here the video, name the chapters, and the transcript, and then it will go ahead, write a blog, and then it will automatically create a new blog entry for me inside of my website.

It embeds the video properly, and it then goes ahead and writes an entire blog post based on my transcript.

And all of that was done without me copying and pasting anything over.

All I had to do is add that video to my custom playlist that I set up inside of Tella.

And so, what used to take me probably, like an hour and a half, maybe a little bit longer after having recorded the video now happens without my input.

And I'm super stoked with it.

So well done, Team Tella on the API.

It's an absolute amazing addition to the platform, and I hope everyone has as much fun with it as I do.

Bye.

So that was a quick look at what you can do with our API.

But today we're also announcing an MCP server.

Michiel, what's an MCP?

An MCP server is a way to extend an LLM agent, like Claude Code, like Cursor, like ChatGPT, so it more easily can connect to external services like Tella.

it can do all of the same things you can do with our public API, like list videos, list playlists and do all of those other actions, but in your LLM agent.

and I'm going to show you a quick demo of this.

So for example, you can use this to organize your videos in a way that is not really easy to do through our UI.

So for example, get my last 10 videos, analyze their transcripts and suggest new playlists.

And so now this is going to make a call to our API through an MCP.

here it's grabbing the last 10 videos, it's grabbing the transcripts for all of those last 10 videos and then it's hopefully going to analyze these and give us new playlist IDs.

And so these playlists aren't playlists that you've already got in your account.

You're asking for, like come up with new ideas for how to organize these ideas.

Exactly.

Cool.

it's looked at all of the videos.

so the last 10 videos were the 12 Days of Tella and some internal videos.

So suggested playlists that it did were Tella Editor, masterclass.

So we can put some videos in there, Quick Mode, Internal Tests, Private and Tella for Teams.

So I guess what's cool about this is that you can kind of just explain what you want to do with a bunch of new videos, get new ideas for it, and then let Claude kind of do all of that instead of having to figure out how you're going to do that in the UI.

Yeah, exactly.

And one other thing you can do, for example, is, grab all of the videos in a specific playlist and make YouTube description for it.

So that's also something we could do, like grab all the videos in the 12 Days of Tella, playlist, make a YouTube description for them, and give me the 4K thumbnail.

So last week, when we had the first batch of 12 days of Tella videos, done, I then needed to publish them to YouTube in a batch because I didn't do them day by day.

So this would have been perfect where I could have had the whole playlist and saying, look, I'm publishing these all to YouTube.

Now go and make me a description for each of them and, give me the thumbnail I can use to start making video.

My own YouTube thumbnails.

it has made a description for each video.

So.

So let's start with the new comments.

One.

I can easily copy paste this now to YouTube.

And it's also included, the 4K thumbnail, although it says it's 1080p, which is funny.

But, I can copy this and reuse it in any place I want.

So these were two examples of what you can do.

But, to me, the more interesting part is how you can hook this up into your own workflow.

you can combine, this with other MCPs, like every time you record a video, you can also update a ticket in Linear, send, somebody an email, all of the things that, you want to do.

the last thing that we want to show you today is our new embedded player SDK.

What this does is it lets you have more control over the embedded player wherever you embed your video.

So if you're embedding your videos in a separate course platform, then you can extend the player to integrate better with the rest of the webpage that it's embedded in.

I'm going to hand it over to Ankita, another one of our customers who has already started building with the embed SDK.

Hey everyone.

So I have created this custom course platform on top of this technology called as Next JS.

I have over 300 lessons.

So I'm basically recording all of those lessons in Tella and embedding it directly in my platform.

Now, thanks to the API, I was able to generate these list of chapters.

For example, if you press, it skips to a specific timestamp.

it skips to, 36 second, 149 and so on.

Now all this data is in fact retrieved from the API.

I was also able to retrieve the transcript and display it this way.

If the student wants to skim through it, it's great.

Plus it's also really good for SEO as well.

At the same time, how long it takes to complete the lesson, number of views, number of chapters, and so on.

Now, if this specific lesson, what is Next JS has multiple Tella players, that means multiple videos.

Then I could also just sum it up as well.

So it's all great.

And if you go to the API reference, I'm specifically calling the Get Video Information API that gets all the video information, such as the chapters, how long is the video and description, and so on.

So there's a lot to it.

if we take a look at the lesson, this is where I embed the player and I pass the video id.

I have this component where I pass the video ID and I get the video data.

I call the API here and then transform the data.

This way, only the data can that I need will be transformed to the component I want.

And then this calls the teleplayer.

And this is the component that you see the video player, and it in fact uses the iframe to render it.

An iframe uses the JavaScript API called as PostMessage that allows us to post specific commands that the Tella player can accept.

So some of those commands are play, seek, mute, unmute, set, volume, for example.

the way I generate into chapters is by using the seek command and I send the command, I pass the time after I retrieve that information from the API, and then I simply call the iframe and post message.

It's as simple as that.

the rest of the information, such as a transcript, such as metadata, and all of that can simply be displayed inside the component.

And this is where I render the chapters.

Then this is where if I check if the player is ready.

So I manage the different states of the player, how long is the video, and so on.

So you could do a lot with it, but that's how all of this works.

So that was a quick look at our, embed SDK.

And that wraps things up for day eight of 12 Days of Tella.

Go and check out the API docs, the embed SDK, the MCP.

Let us know what you think.

Send us a message for how we can extend this further and improve it.

we are all ears.

Thanks for tuning in.

And we will see you tomorrow, Friday, for day nine of 12 Days of Tella.

Bye for now.