# Awesome Agentic AI (ZH) — 7-Stage Learning Map

> awesome-agentic-ai-zh is a bilingual 7-stage roadmap for AI agents, with 145+ curated projects, hands-on exercises, and CLI/MCP ecosystem guidance.

## Install

Paste the prompt below into your AI tool:

## Quick Use

1. Read online or clone:
   ```bash
   git clone https://github.com/WenyuChiou/awesome-agentic-ai-zh.git
   cd awesome-agentic-ai-zh
   ```
2. Start here:
   - `resources/setup-guide.md` (30–45 min)
   - `stages/00-foundations.md` → Stage 1 → Stage 2
3. Pick a track:
   - Track A: CLI Power User
   - Track B: Agent Builder

## Intro

awesome-agentic-ai-zh is a bilingual 7-stage roadmap for AI agents, with 145+ curated projects, hands-on exercises, and CLI/MCP ecosystem guidance.

- **Best for:** learners who want a structured path from LLM basics to multi-agent systems and MCP
- **Works with:** Markdown-first content; ZH (Traditional/Simplified) + English; includes exercises and project picks
- **Setup time:** 30–45 minutes to start

## Practical Notes

- Quant: the roadmap is organized into **7 stages** with two tracks (CLI Power User vs Agent Builder).
- Quant: it highlights **145+ projects** and estimates **14–19 weeks** minimum (realistically 5–6 months at 5–8 hr/week).

## Main

A practical way to use this repo at work:

1. Follow Stage 0–2 as shared foundation, then choose Track A (use CLI agents) or Track B (build agents).
2. Treat each stage’s “hands-on” items as acceptance tests: finish them before moving on.
3. Keep a weekly log of what you built and what failed—agent work improves fastest when failures are explicit.

If your goal is productivity with Claude Code/Codex, Track A is the faster path; Track B is the deeper path for system builders.

### FAQ

**Q: Is it only in Chinese?**
A: No—README shows Traditional Chinese, Simplified Chinese, and English docs.

**Q: Which track should I start with?**
A: Most people start with Track A for CLI productivity, then switch to Track B for deeper agent building.

**Q: How long does it take?**
A: The repo estimates 14–19 weeks minimum; realistic pace is 5–6 months at 5–8 hours/week.

## Source & Thanks

> Source: https://github.com/WenyuChiou/awesome-agentic-ai-zh
> License: MIT
> GitHub stars: 1,043 · forks: 102

---

<!-- ZH -->

## 快速使用

1. 在线阅读或克隆：
   ```bash
   git clone https://github.com/WenyuChiou/awesome-agentic-ai-zh.git
   cd awesome-agentic-ai-zh
   ```
2. 从这里开始：
   - `resources/setup-guide.md`（30–45 分钟）
   - `stages/00-foundations.md` → Stage 1 → Stage 2
3. 选择路线：
   - Track A：CLI Power User
   - Track B：Agent Builder

## 简介

这份 AI Agent 中文学习地图把资料按 7 阶段整理，包含 145+ 精选项目与每阶段 1-5 个动手练习，并提供 Track A/B 两条路线与繁中/简中/英文对照，同时给出 14-19 周（现实 5-6 个月）学习时程估算。

- **适合谁：** 希望从 LLM 基础系统学到多 agent 与 MCP 的学习者
- **可搭配：** 以 Markdown 为主；繁中/简中/英文；包含练习与项目推荐
- **准备时间：** 30–45 分钟起步

## 实战建议

- 量化信息：路线分为 **7 个阶段**，并提供两条 Track（CLI Power User 与 Agent Builder）。
- 量化信息：精选 **145+ 项目**，并给出主干 **14–19 周**（现实 5–6 个月、每周 5–8 小时）的估算。

## 主要内容

更适合落地的用法：

1. 先跑完 Stage 0–2 的共同基础，再按目标选 Track A（会用 CLI agent）或 Track B（从零构建 agent）。
2. 把每阶段的“动手练习”当验收标准：做完再进入下一阶段。
3. 建议做周度学习日志：记录做了什么、卡在哪里、怎么解决；agent 相关技能的进步离不开显式复盘。

如果目标是尽快用好 Claude Code/Codex 提升效率，优先走 Track A；要做系统与编排能力则走 Track B。

### FAQ

**只有中文吗？**
答：不是；README 显示有繁中、简中和英文对照版本。

**先走哪条路线？**
答：多数人先走 Track A 把 CLI 用起来，再回到 Track B 学系统构建更高效。

**大概需要多久？**
答：仓库给出的估算是主干 14–19 周；现实通常是每周 5–8 小时持续 5–6 个月。

## 来源与感谢

> Source: https://github.com/WenyuChiou/awesome-agentic-ai-zh
> License: MIT
> GitHub stars: 1,043 · forks: 102


---
Source: https://tokrepo.com/en/workflows/awesome-agentic-ai-zh-7-stage-learning-map
Author: Prompt Lab