This is my second app built on bubble. It’s called Scribr and is affordable transcription that uses machine learning to automatically summarize your transcript into the important parts. It’s targeted at the likes of journalists, students and lawyers.
It’s more mobile friendly than desktop, and you get your first transcript free, so give it a go!
It uses two different APIs to do the transcription and summarization, but aside from that, all built on Bubble.
I’ll have an Android app out next week, which I’ll put together more of a guide on, because I actually built that using Jasonette for a fully native experience and used my Bubble workflow and data APIs as the backend. It is actually really cool, and was nowhere near as hard as I thought it would be.
But please check out the web app for now and let me know what you think!
Check out the pricing link at the bottom. Looks like it’s for another product.
Oh whoops. Thanks for that! I copied that footer from my other site
Your app looks amazing, @harrytuckerr
Would you be willing to share some of your process of building the app out in Bubble and then using Jasonette to create a more native experience? I’d love to even see some resources on how you learned to do this!
Definitely! I’m a bit smashed trying to monetise it this week. But I’ll try do something in the next few weeks when I finally finish the Jasonette to get in more detail.
But on a super high level for now, I essentially created a stacked single page group for mobile like many people suggest for the mobile experience. I designed it with a mobile app UI thought process from day 1 on the mobile side. I also added the manifest file and allowed it to be used as a progressive web app for now to make it more native.
Now with Jasonette, I’m almost done, just a couple screens to go now. It’s quite easy once you get the hang of it. On the docs part if you just follow the view section step by step you pick it up pretty quickly. I would recommend adding your calls to your Bubble API database in the header so it’s easy to keep calling throughout.
Also, a massive time saver was using these code snippets and writing it in Visual Studio Code.
Pretty cool Harry. Which API’s did you end up using. There are a lot of tools that claim to be able to do transcribing using machine learning / AI but I think this is the first one I’ve seen that is purpose built for it.
I actually combined and built a few of my own tools with it. At it’s core it’s using Amazon Transcribe, but then we have our own neural network for natural language processing in TensorFlow that we combined it with to help with accuracy for certain topics, as well as summarize the key points. I packaged that all together and hosted it on AWS Lambda and AWS’ API Gateway so it had REST API points to use with Bubble.
Very cool, Harry!
I think this is still a bit out of my realm, as I have no previous experience coding. However, the insight is much appreciated!
I actually had no real experience properly coding 18 months ago either, but through using Bubble it taught me without realising! When I first started picking up and trying to code 6ish months ago I found it a breeze to learn, as Bubble’s workflows and how they make you think are actually really similar to how most web languages work, it’s just a matter of remember syntax after that. I highly recommend trying to learn some code if possible, it’s allowed me to create whatever I wanted without much external help and in record time using Bubble for about 80% of it.