Tally

Over the weekend, I built a new app to help me track the food I eat — it’s called Tally.

What is Tally?

Tally makes it easy to track what you’re eating by simply writing it out. For example, writing “1 breakfast sandwich (3 large eggs, 1 tsp butter, 1 slice sharp cheddar cheese, 1 jalapeno, and 2 slices Canyon GF bread)” will return a table with all the macro data broken out, ready for you to save and add to your totals for the day.

Here’s a look:

Tally provides a smart set of default macro settings out of the box, and you can easily customize them to meet your own goals.

Additionally, you can manually log your weight, or, if you use a Withings scale, connect it to your account so the data syncs automatically.

Finally, you can connect your Strava account to incorporate the calories you burn from your runs, rides, etc. in your daily totals.

Why did I build it?

I tried tracking my diet a few years ago, but found it time-consuming. There were a few promising apps, but they fell short in terms of user experience. It sucks to have to type out each ingredient, search for it in their database, specify the amount, etc, etc. All of my previous good intentions failed.

About a month ago, I started using ChatGPT and then Claude to track what I was eating. I would type out what I was eating, and each app would keep a tally for the day and offer recommendations on what to eat next, and pitfalls to avoid. Super helpful! There were issues, though. At times, food would get added to the previous day’s totals. There were also times when I wanted to add something I had forgotten, and these situations felt clumsy. I also couldn’t get a sense for day-over-day trends. Finally, if I worked out, I’d have to add that data by taking screenshots in Strava and pasting them in a chat session. None of that was a dealbreaker, though.

Then, last Friday, Claude asked me if I wanted to build my own app. I replied that I had no interest, but the more I thought about it, the more fun it seemed to build. A few hours later, I deployed a simple v1 Rails app to Heroku and was thrilled to start using it.

On Saturday morning, while waiting for my son’s hockey team to warm up, I added an integration with Withings, the electronic scale I use to weigh in at home.

On Sunday morning, I did the same with Strava. You can see how activities are incorporated into the calorie count in the first screenshot above.

What’s next?

I plan to add the ability to post a photo of what you’re eating (versus typing it out). I’d like to port this to an iOS app, too. Finally, I’d like to compare the different foundation models and the macro data they produce.

Stepping back, it is just ridiculously easy to build right now. The Anthropic API made parsing this type of data dead simple, and of course, the Claude models are so efficient in helping you build. There has never been a better time to build, and paradoxically, there will never be a worse time to build than right now (the pace of innovation in software development right now is absolutely bananas).

If this app looks helpful to you, give it a shot and let me know what you think (jeff@judge.io) - thanks!

Update - March 24

Roughly three weeks later, I’ve continued using this app and loving it. Here’s an insights view I added a few days ago:

This helps me see that I’ve done a good job eating within the caloric limits I’ve set (which are adjusted higher on the days that I work out - the orange ticks), but also that I’m not hitting my protein targets. The higher protein targets are set to help me build more muscle while continuing to lose weight. A trend I can take action on.

I also integrated Withings sleep data to see how it can be leveraged over time.

Naps are displayed too:

Update - April 6

Tally now tracks caffeine consumption:

To track, log your espressos, coffees, etc., the same way you’d log what you eat.

I’ve also added a mood log:

My hope here is to correlate how I’m feeling with diet, exercise, and sleep over time.

Finally, there’s now a feed that provides a helpful day-over-day roll-up: