# Using it without writing code (/en/docs/guide/guides/for-non-developers)

What x-iztro does, where it is typically used, what to hand your engineers, and how accurate it is.



*For: Zi Wei enthusiasts · product and decision makers. No code required anywhere on this page*

## What this is [#what-this-is]

x-iztro is a **charting engine**: give it a date of birth, an hour of birth and a gender, and it
computes a complete Zi Wei Dou Shu chart — and can turn that chart into text a language model can
read.

It is not an app and not a website. It is code for programs to use, and an engineer has to build it
into your own product.

## What it can do [#what-it-can-do]

* **Compute a complete natal chart**: where the twelve palaces sit and their stems and branches,
  which stars fall into each palace, star brightness and mutagens, the four groups of twelve gods,
  the decadals and the age fortune track.
* **Compute the horoscope for any moment**: six scopes — decadal, age fortune, yearly, monthly,
  daily, hourly.
* **Turn a chart into AI-readable text in one call**: no need to describe the chart yourself; paste
  the generated text into a language model and start asking.
* **Six chart languages**: Simplified Chinese, Traditional Chinese, English, Japanese, Korean,
  Vietnamese.
* **Switchable schools**: the year boundary, late Zi hour attribution, the Zhongzhou school and the
  heaven/earth/human plates are all configuration.

What it does **not** do is interpret. It can tell you which star landed in which palace; "how will
this person's career go" is not the library's job — that step goes to an AI or to a person.

## Where it is typically used [#where-it-is-typically-used]

### An AI reading bot [#an-ai-reading-bot]

The most common shape. A user gives their birth details in conversation, the backend calls x-iztro
to chart, and the generated chart text is sent to a language model along with the user's question;
the model produces the reading.

The library gets the chart right, the model does the talking — two separate jobs, each reliable on
its own terms.

### The backend of a Zi Wei app [#the-backend-of-a-zi-wei-app]

The app's frontend draws the chart and handles interaction; the charting computation sits on the
server. One body of computation can serve iOS, Android, web and mini-programs at the same time.

### Bulk data analysis [#bulk-data-analysis]

Questions like "across these hundred thousand birth records, what fraction have Ziwei in the Soul
palace?" or "how does a given pattern correlate with a given field". x-iztro charts in milliseconds,
so hundreds of thousands of records finish on a single machine.

## Do we need an engineer? [#do-we-need-an-engineer]

Yes. x-iztro is a code library, not software you can open.

But the work is small: install it, write three lines of calls, wire the result into a flow you
already have. A competent backend engineer usually has a working prototype inside half a day.

## What to tell your engineers [#what-to-tell-your-engineers]

Passing along the following is enough:

| Item                 | Detail                                                                                            |
| -------------------- | ------------------------------------------------------------------------------------------------- |
| Repository           | [github.com/x-haose/x-iztro](https://github.com/x-haose/x-iztro)                                  |
| Programming language | Rust, Python or Go — pick one, the results are identical                                          |
| Installation         | Python: `pip install x-iztro`; Go: `go get`; Rust: `cargo add`                                    |
| Requirements         | Python 3.10 or later / Go 1.22 or later / Rust edition 2024                                       |
| Inputs               | Three things: date of birth, hour index (0–12), gender                                            |
| For AI use           | Call `chart.to_text()` to get the chart text and feed it straight to a model                      |
| Docs                 | [Getting started](/en/docs/guide/getting-started), [Semantic text](/en/docs/guide/guides/to-text) |

<Callout title="How to fill in the hour index">
  It is not the clock hour; it is one of thirteen values, 0 through 12. 0 is the early Zi hour
  (00:00–01:00), 1 is Chou, 2 is Yin, … 11 is Hai, and 12 is the late Zi hour (23:00–24:00). The full
  table is on [The inputs you need](/en/docs/guide/getting-started#hour-index).

  **Born after 23:00 means 12, not 0** — the two values produce different charts.
</Callout>

## How accurate is it? [#how-accurate-is-it]

"Accurate" has a very specific definition here: &#x2A;*identical field for field to the JavaScript
[iztro](https://github.com/SylarLong/iztro)**.

iztro is one of the most complete and longest-maintained open-source Zi Wei Dou Shu charting
implementations, and a fair number of frontend projects use it. x-iztro treats it as the reference
and checks against it field by field across 716,314 test cases — any divergence is treated
as a bug and fixed.

The **boundary** of that claim matters too: agreeing with iztro is not the same as being "the one
correct answer in the art". Zi Wei has many schools, and different schools genuinely place stars and
assign mutagens differently. What x-iztro guarantees is "given the same set of school choices, it
computes exactly what the reference implementation computes" — not "this set of school choices is
the right one".

See [Accuracy](/en/docs/guide/about/accuracy).

## What MIT means [#what-mit-means]

x-iztro is open source under the MIT licence. For a user, that means:

* **Commercial use is fine**, with no fee and no need to notify the author.
* **Closed-source use is fine**: build it into your commercial product and the product itself need
  not be open sourced.
* **Modification is fine.**
* The single obligation is to **keep the copyright notice** (usually on the product's "open source
  licences" page).
* The author **provides no warranty**: if something goes wrong in use, that is on you.

Among open-source licences, MIT is the least restrictive category, and legal teams rarely object.
