Best AI saju & BaZi readings in English (2026) — an honest comparison

What actually separates these services

Every AI reading service does the same two jobs: calculate the chart, then interpret it. The interpretation is taste — warm or formal, long or short. The calculation is not taste; it is either right or wrong, and most marketing carefully avoids the topic. So compare on four things: whether the math is transparent and reproducible, how generous the free tier is, whether pricing is one-time or a subscription, and how refunds work.

ioreum — one-time credits, engine-first (that’s us)

ioreum is a Korean saju take on BaZi. Its defining choice is architectural: a deterministic engine computes every number — pillars with solar-term boundaries, true solar time from your birthplace, element balance, luck cycles, a 0–100 compatibility score with stated reasons — and the AI only writes the prose. It cannot change a digit, so the reading can be warm without drifting into flattery.

• Free: engine snapshot with no account, for every product; invite a friend to earn a free reading credit.
• Pricing: one-time credits — $1.99 for one reading, $4.99 for three, $9.99 for ten. No subscription.
• Refunds: 14-day self-serve refund on unused packs, from your account page.
• Best for: readers who care whether the math underneath is real, and who hate subscriptions.

Seoul Saju — the generous free tier

Seoul Saju offers AI saju readings in a dozen languages with an unusually generous free tier — readings without signing up, daily free quick readings, and a first premium analysis free. Paid reports are long-form PDFs in the fifteen-page range. If you want maximum free content before spending anything, it is the strongest option; the trade-off is a credit system where premium readings consume multiple credits.

• Best for: sampling a lot of free AI readings before deciding to pay.

K-MUDANG — classical-sources positioning

K-MUDANG leans on classical legitimacy: named source texts, astronomical ephemeris data, and its own calculation-versus-interpretation split. Credit packs run roughly seven dollars for three readings to thirty-eight for twenty, landing near two dollars a reading. It offers saju, multi-category compatibility and face reading. If citations to classical texts matter to you, it is the most explicit about them.

• Best for: readers who want classical-text grounding spelled out.

Cantian AI — the app-and-subscription route

Cantian AI grew from a popular GPT-store calculator into an app with BaZi, name analysis and I-Ching, and claims a very large user base. Its current model centers on subscriptions in the thirteen-to-twenty-dollars-a-month range on top of a free calculator. If you want an ongoing app relationship rather than a one-off report — and don’t mind a recurring charge — it is the established name.

• Best for: people who want an app they open weekly, and accept subscription pricing.

The free calculators — charts without readings

A cluster of free BaZi calculators (BaZi Lab, OpenFate, FateMaster and others) will compute your pillars at no cost, with varying rigor on the boundary problems — solar terms, the 11 pm hour, true solar time. They are useful for cross-checking a chart. What they mostly don’t give you is the reading itself, which is where every service above charges.

A fair bottom line: if you want the cheapest honest full reading with no strings, that is what we built ioreum for — $1.99, one time, refundable. If you want maximum free sampling, start with Seoul Saju. If you want classical citations, look at K-MUDANG. If you want an app with a subscription, Cantian is the incumbent.