Jan 10
API Design
A new project. Designed the backend for a tennis club app. Members can book courts, sign up for lessons, and manage family subscriptions, all from their phone. We have built secure login that works across multiple devices. Payment system handles card payments, bank transfers, and cash with automatic mismatch detection. Family accounts let one member manage and pay for the whole household. The system auto-calculates bundled pricing and early-bird discounts. For the techies out there we built a Laravel API with JWT auth and multi-device tracking (device bans, token refresh). Versioned architecture: 12 controllers, 40+ endpoints under /api/v1/ with consistent { success, message, data } responses. Next week we will be looking at the UI.
Jan 06
Audio Notes Integration
As we really start to pile in the features into LekeeLekee we have no started the introduction of voice notes, these are publicised comments against a post in voice. Lots of considerations here the most important being compliance and copyright. We dont want people to say anything derogatory or indeed share Taylor Swifts latest release. It works well, in React and React native, all served over a Cloudfront CDN.
Jan 02
Write SQL with human language
For years, accessing data meant either learning SQL or waiting on engineers. AI changes that. You can now ask plain-English questions like “Show customers who signed up last month” and get the correct SQL instantly. This means faster answers, fewer bottlenecks, and better use of engineering time. Teams get accurate queries in seconds, learn SQL naturally by seeing results, and stay independent for everyday data needs. It’s not about replacing experts — it’s about freeing them up. The future of data isn’t writing better queries. It’s asking better questions.
Dec 28
Work Tracker
We have built a desktop app that automatically monitors our desktop activity - tracking which apps and windows we use, capturing keystrokes, and taking periodic screenshots. It auto-detects clients based on pattern matching (URLs, project names) so time/tasks is automatically tagged to the right client. View daily activity breakdowns, time spent per client, and generate social media copy from your work sessions. All data stays local on your machine. version 1.0 is done and works perfectly and allows us to update our clients in realtime about what we are working on. The Stack Framework -Electron (cross-platform desktop app) Frontend -React + TypeScript + Tailwind CSS Backend -Node.js (Electron main process) Database -SQLite (local, via better-sqlite3) AI -Claude API (for activity categorization) Build -Vite + electron-builder
v1.0 LekeeLekee
We have had a busy couple of months on the version 1.0 of our African Social media platform. As Christmas approaches, we are now finessing the UI, the little tweaks that make the product pop, small animations, transitions that will make the app pop. Lots and lots going on with the backend side of things for this project too which has certainly pushed us.
Dec 18
GDPR
Yesterday with a client we were talking about GDPR. What sounds simple on paper quickly becomes retention rules, audit trails and background jobs built purely for compliance. The real cost isn’t fines, it’s engineering time diverted from building better products.
Yes, this QR Code is being tracked.
Thanks for scanning the code. You just clicked the QR on the car and landed on our live werkilo feed. We are recording scans to see if these codes ever get used and to learn what works in the real world. We built the tech behind this and the platform you are viewing it on to show real work as it happens not polished case studies. The same tools workflows and thinking we use here are available to you too.
Dec 09
E2EE for LekeeLekee
End to End Encryption, is the ability for content to be only read by two peers. When you send a message in WhatsApp to your friend, the only two people that can read it our your both. Not even WhatsApp can read, as part of our work in building LekeeLekee we need the same technology for the messaging feature. It took a little time, but is now all done. Even if we wanted too we cant read the messages between each other.
Dec 02
Cache within Cache
Wondered how the likes of X and Facebook can output results so quickly while trawling through billions of records. It's done by using a really heavy caching strategy, (it generates the results once and then serves up multiple time). For LekeeLekee this is a key factor in ensuring we have the output at a speed that our users need. While the client doesnt have the billions to have huge engineering teams, they do have us. We have built a caching system running of Redis, which allows for near instant return of results, >10ms at the moment but we will need to keep changing this as more and more records are generated on the fly.
Nov 14
Small Business Support
Not always do we have the biggest and the shiniest on the go. We do have large commitments to some huge infrastructure, and we always want to ensure where possible others can use. We love supporting small business, including a local Pub. What do we do, we simply manage their website hosting, email, and the integration with ResDiary (reservation software). Its not deeply technical and runs of a plesk instance. It works and it runs. Get in touch if you need some hand holding.
About werkilo
Show everyone what you are working on. So at the very least they know what you do and who you do it for!