Taka Shirasu

Always a great way to start with a small NoSQL document database. MongoDB Atlas allows us to start with a free tier and flexibly adapt as we scale up our monitoring, traffic, and storage requirements. MongoDB also allows us to be quite flexible with our schema definition as our product changes a lot initially.

We picked GitLab over GitHub because we prefer GitLab’s privacy policy which is slightly more transparent about usage of the code in AI tool contexts.

Sometimes we need endpoint handling outside of our Vercel application. We use AWS Lambda with EventBridge Schedules to send email with a delay. AWS Lambda is very cost effective for scaling compute both up and down.

Server components make it much more efficient to reason about server side rendering and using / optimising assets. It also draws a clear line between what is client / server side rendered. The App router makes routing easy and allows for a nice folder structure.

The flexibility of the platform and how easy it was to create a database made us stick to the product.

Since everyone uses this tool, there is many of support when I stumble. Also, FigJam is the perfect tool for product management that I can use alongside the design on the Figma file.

It is easier to start the meeting compared to any other alternatives.

It was easiest to start as most other poeple are using and more stable for translation compared to other LLM API.

Configuring email is usually a headache. And when you just about get it right, it goes into recipients’ spam folder. Resend just takes the pain out of email and just works every time.

We need to minimise the amount of time we spend setting up deployments and optimising our website hosting. Vercel makes it so easy to automatically deploy when code changes are pushed to our repo.