Splync v1.9 Can Set Custom Sharing Ratios for Each Expense!

About one and a half months after releasing Splync v1.5 — the update that finally allowed custom split ratios per project and per category — a new wave of feedback arrived. The v1.5 update demanded heavy server-side changes, so at the time, I believed the granularity was “good enough” for most use cases. Then a few new users asked a simple, very reasonable question: “Can we set custom sharing ratios per individual expense?” Their trip project had moments where only two members needed to split a specific cost, and other moments where all three wanted to share something. Under Splync v1.8, the answer was unfortunately no. I explained that they could create additional categories with custom ratios tailored to those specific cases — a workaround, not a true solution.

It felt a little overwhelming to remember how much work went into v1.5. First I had to let projects have their own split ratios. Then categories needed their own custom ratios too. At that point I thought I had covered the whole tree — gathering every fruit from every branch. But working on per-expense ratios was different. It felt as if each time I picked a fruit, a new tree sprouted from that very spot. Not an infinite fractal forest, but a clear two-story structure: one layer giving birth to the next. In the v1.5 logic, an expense first inherited the project’s ratio. If its category had custom ratios, those would overwrite the project’s values. So when adding per-expense ratios, I found myself trying to add another overwrite on top of that overwrite. The structure became a staircase of overrides — technically correct, but mentally messy. It was hard to justify building another layer of patchwork logic.

The breakthrough eventually came from flipping the structure upside down. Instead of doing “project → category → expense” and overwriting each layer with the next, why not think in the opposite direction? Per-expense shares → per-category shares → per-project shares. That ordering reflects how real people may think: if a specific expense has its own rules, it should simply follow them. If not, the category’s pattern makes sense. If even that fails, fall back to the project default. No more staircase of overrides — just a clean hierarchy of priority. Once I saw the structure this way, the fog of the pseudo-fractal forest cleared instantly. The implementation path became obvious: “Each expense is checked for per-expense custom shares. If present, use them. If not, check per-category shares. If not, use the project default.”

The new logic felt simple, predictable, and mathematically sound. To support this priority system, we added a dedicated MariaDB table for per-expense splits, sort of mirroring the table for per-category splits. The expense details table also needed expansion, just like the category details table did back in v1.5. Once the underlying “two-story” design became clear, the rest was just careful coding — both in the app and on the server. Still, the path felt risky in places, like walking through a dark forest without a map. I wanted to clear my head and stepped outside. The air was sharp and clean. While walking through my neighborhoods, I saw Mt. Fuji rising in the distance, crystal blue beneath a perfectly blue sky. It was almost 100 km away, yet it looked close enough to touch. The moment felt like a reminder: even if I think I’m walking in a forest, I’m actually walking under the open sky.

Splync v1.9 was born from that moment of clarity. I returned home, finished wiring the new endpoints, prepared the new server logic, organized the related interfaces, submitted v1.9 to Apple, and finally fell asleep. The review finished earlier than usual. When I woke up, Splync v1.9 had already been approved and automatically released on the App Store. From this version onward, splitting becomes far more flexible. If you’re traveling with John and Kate, you can share basic expenses evenly among the three of you. But for food, you might switch to a “25% : 50% : 25%” split because John usually eats twice as much. And if John skips one dinner—say, at an oyster restaurant—you can set that specific meal to “50% : 0% : 50%” so he doesn’t pay for what he didn’t eat. With v1.9, Splync now supports per-project, per-category, and per-expense ratios in a unified logic. The foundations are stable and mathematically sound. The next challenge is the interface: some other splitting apps offer smoother, more stylish ways to adjust those ratios. Splync now has the backend strength to support such improvements. We will work on them one after another.