支叶平台

会学习的人机交互内核，及其可教学的语言接口——Y语言

无需重新编译，用对话赋予机器新能力

Gitee 仓库立即体验

从"编程"到"教学"：一种新的交互范式

传统自动化需要为每个新功能重新编码与编译。在支叶平台上，你通过Y语言"教导"系统新指令与新逻辑。平台在运行时理解并记住它们，实现交互能力的持续生长。

核心构成：平台与语言

支叶平台

可学习的交互内核。提供记忆、执行、多模态感知等基础能力，并能在运行时吸收新知识。

Y语言

平台的动态接口。不是需要编译的静态语言，而是可被实时"教学"以扩展平台理解的交互协议。

平台为交互注入的能力

支叶内核提供的底层能力，通过Y语言直接调用

持续学习

在交互中实时记忆指令、函数与规则，系统越用越智能。

硬件操控

直接驱动串口、GPIO等，连接物理世界。已部署于RV1106等设备。

多模态感知

集成视觉(YOLOv5)、语音等感知能力，让机器拥有"感官"。

自然语言理解

解析复杂指令，将其映射为可执行的任务流与函数调用。

支叶融合 — 连接 AI 与万物的底座

一面接入万物（硬件、传感器、设备），一面直连 AI。Y 语言沟通一切，大幅降低行业 AI 门槛，高效无限扩充。

AI 调用支叶

AI 通过 Y 语言直接操控摄像头、串口、UI、语音等全部扩展能力，为 各行各业具体业务 快速赋能。

AI 进化支叶

AI 参照 C++ 扩展模板 自己开发、编译、测试、使用新的 DLL，支叶能力边界随 AI 无限生长。

接入 AI 助手

Claude Code

终端运行 claude，提到支叶或 /nlu 加载知识库

WorkBuddy

打开 nlurh/ 文件夹，提到支叶即自动加载

项目内置 Skill 文件（.claude/skills/ 和 .workbuddy/skills/），从 Gitee 下载后开箱即用，无需配置

如何开始"教学"？

运行平台

启动 _zhiye.exe

用Y语言教学

输入指令或示例

平台学习并反馈

理解、记忆、执行

立即体验

进入 nlu_run_32bit 或 nlu_run_64bit 目录，双击运行 "_zhiye.exe"。使用密码 1234，即可开始用自然语言与系统对话。

资源： 项目文档 · 示例 · 组件集成指南 · 商标指南

Zhiye Platform

A learnable human-computer interaction kernel, and its teachable interface — the Y Language

No recompilation needed. Empower machines through conversation.

GitHub Repository Try It Now

From "Programming" to "Teaching": A New Interaction Paradigm

Traditional automation requires re-coding and recompiling for every new feature. On the Zhiye Platform, you "teach" the system new instructions and logic through the Y Language. The platform understands and remembers them at runtime, enabling continuous growth of interactive capabilities.

Core Components: Platform & Language

Zhiye Platform

The learnable interaction kernel. Provides foundational capabilities like memory, execution, multimodal perception, and can absorb new knowledge during runtime.

Y Language

The platform's dynamic interface. Not a static language that needs compilation, but an interactive protocol that can be "taught" in real-time to extend the platform's understanding.

Capabilities Infused by the Platform

Underlying powers provided by the Zhiye kernel, directly invoked via the Y Language

Continuous Learning

Memorizes instructions, functions, and rules in real-time during interaction. The system grows smarter with use.

Hardware Control

Directly drives UART, GPIO, etc., connecting to the physical world. Deployed on devices like RV1106.

Multimodal Perception

Integrates vision (YOLOv5), speech, and other perceptual abilities, giving machines "senses".

Natural Language Understanding

Parses complex instructions, mapping them into executable task flows and function calls.

Zhiye Convergence — The Bridge Between AI & Everything

One side connects to everything (hardware, sensors, devices), the other side connects directly to AI. Y Language bridges it all — dramatically lowering the barrier for industry AI, infinitely scalable.

AI Calls Zhiye

AI uses Y Language to directly control cameras, serial ports, UI, speech — empowering industry-specific business across every sector.

AI Evolves Zhiye

AI references the C++ extension template to develop, compile, test, and use new DLLs — Zhiye's capability boundary grows infinitely with AI.

Connect Your AI Assistant

Claude Code

Run claude in terminal, mention Zhiye or type /nlu

WorkBuddy

Open nlurh/ folder, mention Zhiye to auto-load

Built-in Skill files (.claude/skills/ & .workbuddy/skills/) — works out of the box after cloning, no configuration needed

How to Start "Teaching"?

Run the Platform

Launch _zhiye.exe

Teach with Y Language

Input instructions or examples

Platform Learns & Responds

Understands, remembers, executes

Try It Now

Navigate to the nlu_run_32bit or nlu_run_64bit directory and double-click "_zhiye.exe". Use the password 1234 to start conversing with the system in natural language.

Resources: Project README & Full Documentation · Example Code (examples/) · Component Integration Guide (docs/dll) · Trademark Usage Guide