When I couldn’t find the right tool for what I needed, I built my own. Here are four of them.
I fix things for a living. I work in cloud infrastructure, mostly AWS. So when parts of my transition got messy, I did what I always do. I built something.
All of these live on my own website, some behind a login I built too. Each one started small, fixing one problem for me, and then grew.
The surgeon finder
When I started looking for a surgeon for facial feminisation surgery (FFS), everything was everywhere. Multiple websites, Google searches, notes from Instagram reels. All of it written on a notepad.
So I built something better. The first version only covered FFS. I used Claude to search the web, several times over, for US-based surgeons who perform FFS. Then I added a ranking system, a call log, and wherever I could find it, World Professional Association for Transgender Health (WPATH) details, insurance details, website links, phone numbers and emails. Everything in one place.
The ranking was simple. Red for no. Orange for maybe, I still have questions. Green for yes, she’s on the shortlist. Only two made it to green. And the funny thing is, the surgeon I chose went straight past green. I didn’t even rate her. I felt that strongly about her. She was gold. She broke the rating system without trying, or even knowing it existed. It was never needed for her anyway.
Then it grew. I rebuilt it to cover every gender-affirming surgery, not just FFS. It now holds 1,115 surgeons across all 50 states and 25 different procedures, pulled from the public directory at transhealthcare.org. That directory only had phone numbers for about 6% of them, so I filled the gaps from the national provider registry and more web searches. Now about 99% have a phone number and about 91% have a website.
You can filter by procedure, state, insurance, and whether they mention WPATH. Every surgeon has quick links to their website, Google Maps and the review sites. The call log grew into a proper tracker: when I contacted them, how, what happened, any quotes, and when to follow up. Once a consult is booked, it shows up in its own appointments list. There are charts too, by state, by procedure, and how my outreach is going. And I can export the lot to a spreadsheet.
It’s good. Really good, in my humble opinion. I want to refine it, and one day I’d like to share it.
The countdown
The countdown started with my first surgery. You name the event, set the date, and it counts down the days, weeks and weekends until it arrives. I’ve used it for other surgeries since, but it works for anything you’re counting down to.
It’s inspired by old-school rip sheets. You write or print a number on each page, one per sheet, counting down. Staple them together, then rip one off each day until you reach the big day.
Mine doesn’t look like paper. It started with one date and now holds several, each with its own colour, so I can watch more than one big day coming. Across the top are big cards showing days, weeks and weekends to go. Below that is a table with a row for every day. Weekends are shaded, today is highlighted, and each event’s zero day lights up in its colour. It changes every day at midnight, and past days just drop off.
The feeling is the same as the paper version, though. One less day to wait.
The gallery
I built my own gallery because I wanted control. Some galleries are public, so anyone can see them. Others are private, and only the people I add can get in. Each gallery can have its own sections, and it handles videos as well as photos.
Some photos are for work colleagues. Some are for my health providers. Some for friends, some for family. And then there are different levels of family.
I can also share my data from any of these tools with other people, as many as I like. Read-only, or full access. My choice, every time.
The nutrition tracker
This is the big one. It started as a simple food log and turned into the most complicated thing I’ve built for myself.
Logging food. Every meal goes in under breakfast, lunch, dinner or snacks. I search the USDA food database and it fills in the numbers for me, or I pick from my own saved foods. It tracks calories, protein, fibre, carbs, sugar and net carbs, and it records the time I ate. That turned out to matter more than anything else. If I’m guessing at a label, I tick “rough guess” so the app knows not to fully trust that entry.
Getting the numbers right took some work. At one point packaged foods were logging about three and a half times too high, because the database reports everything per 100g but the label serving might be 28g. Fixed now.
Goals and limits. Protein and fibre have targets to hit. Carbs, net carbs and sugar have limits to stay under. A 30-day chart colours each day green or red against those limits.
My CGM. I wear a continuous glucose monitor (CGM), a FreeStyle Libre 3, on my arm. Every 15 minutes a small job on AWS pulls my latest glucose readings and stores them alongside my food. The day view plots my glucose curve with a marker at every meal, so I can see exactly what happened after I ate. If my glucose drops low, a warning banner shows up. The Libre app’s own alarm is still the real safety net, though.
Meal spikes. This is the clever bit. For every meal, the app finds my glucose just before I ate, the highest point in the two hours after, how far it rose, and how long it took to come back down. Each meal is rated good, moderate or poor. Meals too close to another meal, or where my glucose was already falling, get left out so they don’t muddy the numbers.
Then it looks back over the last four weeks. It shows how my rise changes with the amount of carbs, and which foods are the worst offenders and which ones behave. When a meal has several foods in it, the rise gets split between them based on their share of the carbs. That shows which food is probably to blame, though it can’t prove it.
Actions. I can log things I do around a meal, like a walk or a glass of vinegar water. The app then compares meals with and without them. It won’t give a verdict until it has at least ten meals on each side, because I don’t want it fooling me with small numbers.
Food scores. Every food gets a blood sugar score from 0 to 100, plus star ratings for weight loss and muscle gain. These are estimates worked out from the macros, not a lab-tested glycaemic index. Each day also gets a quality score compared with the last week.
Medications. I log my meds too, with a pill-organiser style button for each part of the day. Every change in dose gets a date, so the app can compare meals within the same treatment and not mix them up. It only keeps records. It never suggests changes.
And the rest. Weight and BMI, body measurements, weekly and monthly charts, averages and a day streak.
Under the hood
For the techies: it all runs on AWS. The site sits on S3 and CloudFront, logins go through Cognito, and the back end is Lambda and API Gateway, with DynamoDB for the data and S3 for the photos. The nutrition tracker adds a scheduled job that pulls my glucose every 15 minutes. I built it all with Claude Code.
Still building
Some of these will stay mine. Some I hope to share one day. Either way, I’m still building.