Founder note · Building in public

What I Learned Building a BaZi Tool as a Chinese Maker

A Chinese BaZi enthusiast shares five product lessons from community research, chart accuracy, unknown birth time, privacy and visual design.

BaZi is easy to name and hard to enter

I grew up Chinese, and BaZi was one of those ideas that always seemed culturally familiar without being easy to explain. I knew the words. I had seen the eight Chinese characters arranged in a chart. But knowing that something belongs to your culture is not the same as understanding how it works.

That made me exactly the kind of person I wanted to build for: curious, willing to learn, but not ready to decode a wall of traditional terminology. I am a BaZi enthusiast, not a professional practitioner. My goal was not to replace an expert reading. It was to make a first encounter with the Four Pillars of Destiny feel clear enough that someone could keep exploring.

BaZi is one specific tradition within the broader world often described in English as Chinese astrology. It goes beyond a single zodiac animal by looking at the year, month, day and—when available—hour of birth. Yet most English-language introductions still start at one of two extremes: a zodiac animal with little context, or a chart full of stems, branches, hidden relationships and unfamiliar labels. The interesting middle was missing: a BaZi calculator that shows the real structure, explains the first layer in plain English and lets people have fun without pretending every line is a prediction.

The questions I kept finding

Before building, I read public Reddit discussions, compared existing calculators and noted the questions that appeared around chart tools. This was qualitative research, not a survey, and it does not tell me how common each concern is. It did reveal where a first experience tends to break down.

People could generate a BaZi chart but still had no idea where to begin. Some were unsure whether an AI explanation was interpreting a correctly calculated chart or quietly inventing the chart itself. Others did not know what to enter when their birth time was missing. Privacy also mattered because birth date, time and place feel more personal than the average calculator input. And many beginners arrived expecting only a Chinese zodiac sign, not a Year, Month, Day and Hour structure.

Those are not copywriting problems alone. Each question changes the product. A tooltip cannot repair an unreliable chart, and a beautiful result cannot make an invented Hour Pillar honest.

What the research changed

I turned each recurring question into a product rule. The table below is the shortest version of the build brief I wish I had when I started.

What I foundWhat changed in the product
A chart looks like a code sheet to a beginner.Keep the real Four Pillars visible, then label the Day Master, Five Elements and relationships in plain English.
An AI reading can sound convincing even when the foundation is unclear.Calculate the BaZi chart first with calendar rules; add the explanation only after the chart exists.
Many people do not know their exact birth time.Show the three pillars that can be calculated reliably and mark the Hour Pillar as unknown.
Birth details create a real privacy hesitation.Make the calculator free with no sign-up, and do not create a public personal-result URL.
BaZi is often confused with a zodiac animal.Explain that the zodiac is one branch of the Year Pillar, while BaZi reads the full birth moment.

Why chart calculation comes before AI

This was the most important technical boundary. A BaZi reading has two different jobs: form the chart, then interpret it. The first job follows calendar, seasonal and location rules. The second turns relationships in that chart into language. Mixing those jobs makes it hard for a user to know what is calculated and what is interpretive.

So the calculator forms the pillars before any AI-written explanation is requested. The result keeps the chart visible even if a richer reading is unavailable. That sounds like an implementation detail, but it is really a trust decision: fluent language should never be allowed to hide the foundation it is describing.

The same boundary shapes the educational pages. If someone wants to understand a term rather than accept a paragraph at face value, they can read what a Day Master is, how the Five Elements relate or how a BaZi chart is organized. The explanation should open the system, not ask for blind trust.

Unknown birth time should produce less, not fiction

The Hour Pillar depends on birth time. When that time is unknown, many products face a tempting UX shortcut: ask for a guess so the result looks complete. I chose the less dramatic path. The calculator returns the Year, Month and Day pillars and clearly marks the missing Hour Pillar.

A partial result is still useful. The Day Master comes from the Day Pillar, and the visible chart can still introduce the core structure. What changes is the confidence boundary. The tool should say what it knows and stop where the input stops.

That principle also keeps the product welcoming. Not knowing an exact birth time should not block someone from learning the basics, but it should change what the interface claims to know.

A Chinese visual language without costume

I also wanted the tool to feel Chinese without turning culture into decoration. It would have been easy to add dragons, gold gradients, bright red panels and a mystical fortune-telling theme. None of those choices would help someone read the chart.

Instead, the interface takes its structure from the subject itself. The four-column matrix echoes the Four Pillars. Chinese characters remain visually important rather than becoming tiny annotations. Warm paper gives the page softness, deep jade provides calm contrast, and restrained cinnabar works like a seal instead of a warning color.

For me, Chinese aesthetics here means proportion, rhythm, empty space and material feeling. It is less about adding symbols and more about letting the system's own visual logic lead. The result can still feel contemporary, because cultural specificity and modern usability are not opposites.

How the first public version shipped in one day

The first public version went from a written product direction to the official domain on the same day. That did not happen because the product had every possible feature. It happened because I kept the boundaries small and made the risky decisions early.

The core journey was one page: enter birth details, calculate the chart, read the first layer and continue into learning material if a term was unfamiliar. There was no account system, subscription, dashboard or content management system. Free and no sign-up were not temporary launch copy; they removed friction from the exact moment when a curious person decides whether the subject is worth exploring.

The hard parts were not the number of screens. They were deciding what the tool could claim, validating the chart path, handling an unknown time honestly and keeping personal inputs out of public URLs. Once those rules were settled, a small product could move quickly without becoming careless.

A learning tool, a reflection tool and a little fun

BaZi is a traditional symbolic system, not a scientific test or a machine for fixing someone's fate. Interpretations vary between schools, and no chart should replace practical judgment. I think the tool is most useful when it helps someone ask a better question, notice a pattern or understand why the chart is arranged the way it is.

It can also simply be interesting. You can compare the Four Pillars with the single zodiac sign you already knew, find your Day Master, or see how Wood, Fire, Earth, Metal and Water appear in the visible characters. Learning does not have to feel like homework, and having fun does not require making exaggerated promises.

That is the balance I am still working on: enough structure to respect the tradition, enough explanation to welcome a beginner, and enough restraint to leave room for curiosity.

Research notes

Representative public discussions

These threads are examples, not a representative sample or evidence of how common each concern is. I used them as design prompts alongside calculator comparisons; the product decisions are my own.