Web article to audio podcast generator

September 2023 • Astro, SQLite, Bullmq, Azure TTS • GitHub repo

Summary:

Web app that would allow users to save online articles they want to read.
It generates an audio version of the article using the Azure text to speech API, and then give them a podcast feed so they could listen to the articles using their podcast player app.
This uses the Astro framework, SQLite for the database and Bullmq for the background task queue.

Why I made this project

I’ve never been a big reader - I struggle to maintain my attention when reading longform text content. In university when I majored in History, I found it really difficult to get through my readings each week.
I’m a big fan of podcasts and audiobooks - I find it much easier to pay attention and retain information when I listen to it.
In the last couple of years, AI text to speech services have improved in a huge way. I’m a big fan of Microsoft’s Azure Text to Speech API, OpenAI’s text to speech API and Eleven Labs.
In my work at the Information Access Group, I was always interested in generating alternative versions of content. For example, generating audio versions of documents. You can find out more in my writeup about the Easy Read HTML product.
Because it’s relatively simple to make a podcast by generating an RSS feed, I thought it would be interesting to make an app that generates a personalised podcast feed of articles you want to read. That way, you can listen to articles in your preferred podcast player.

I started making a basic web app with Astro, connected with HTML forms.
Using the better-sqlite3 npm library, I added a SQLite database. The application code uses simple SQL to interact with the database, rather than using an ORM like Prisma or Drizzle.
For the background task queue, I used

I learned a lot about SQL and SQLite - how to get a basic database up and running, and how to interact with it using the better-sqlite3 library. Before this project, I was hesitant to build a database-backed web app without using an ORM. But thanks to online resources and tools like ChatGPT and Copilot, writing simple SQL queries feels pretty straightforward to me.
I learned more about handling background tasks with BullMQ and Redis. By building a simple task queue for handling the audio generation, I gained a better understanding of how Redis works and its use cases.

I would like to try rebuilding this web app using a different software stack. While Astro + SQLite is a good combination for simple web apps, I would like to try building it using a more mature framework like Laravel. This would make it easier to turn this into something real, and maybe a product that I could charge a subscription for.