Install Kinogaki Core by hand

The friendly path is to let a coding agent install Core for you: point it at agent.kinogaki.com. This page is for when you'd rather do it by hand, with every step spelled out.

What you're getting

Core ships as a universal macOS static library (arm64 + x86_64) plus its headers, in one .tar.gz. A static library is a compiler input that links into your binary, so it ships as plain object code: macOS Gatekeeper and notarization apply to launchable apps, leaving the archive ready to link as is. Core stands on the C++ standard library alone, with the codecs built in. The current version is 0.1.0.

Download and verify

curl -fsSLO https://cdn.kinogaki.com/kinogaki-core/0.1.0/kinogaki-core-0.1.0-macos.tar.gz
echo "b3ca00d397fc1fca162f7970b0daa7867bde6058d5eb9695789e13baf78eab35  kinogaki-core-0.1.0-macos.tar.gz" | shasum -a 256 -c -
tar -xzf kinogaki-core-0.1.0-macos.tar.gz

The archive unpacks to:

kinogaki-core/
  include/kinogaki/…        the headers: Document.h, Codecs.h, Serialize.h, …
  lib/libkinogaki-core.a    the prebuilt universal static library

Compile against it

Any C++20 compiler works; add the include path and link the archive:

c++ -std=c++20 -Ikinogaki-core/include app.cpp -Lkinogaki-core/lib -lkinogaki-core -o app

Everything is under the kinogaki:: namespace and the kinogaki/ include root. A one-file check:

#include "kinogaki/Document.h"
#include <cstdio>
using namespace kinogaki;
int main() {
    Document doc;
    doc.append(Path("/hi"), "note").set("msg", Str("it works"));
    std::printf("%s", doc.toString().c_str());
}

Build from source instead

Useful on other platforms, or to track main:

git clone https://github.com/kinogaki/kinogaki-core
cd kinogaki-core && ./build.sh

build.sh is a single shell script that compiles the library and runs the test suite.

Python

The package bundles the native library, so one command installs everything:

pip install kinogaki

import kinogaki as kg
doc = kg.Document()
doc.load("notes.md")          # normalize a Markdown file into the document model
print(doc.to_string())        # the .prisma text