Skip to content

Releases: bestdeejay-design/awesome-ai-handbook

v1.0.1 — Emoji→SVG, level indicators, catalog navigation

Choose a tag to compare

@bestdeejay-design bestdeejay-design released this 27 Jul 09:31

🎯 What's Changed

  • Emoji→SVG replacement: All section emoji icons replaced with Lucide SVGs (20×20) across all 92+ content files — consistent crisp rendering everywhere
  • Level indicators redone: Inline SVG circles → <img> tags referencing local assets/icons/level-*.svg files with absolute GitHub Pages URLs (12×12)
  • Code blocks preserved: All emojis inside fenced code blocks left untouched
  • Catalog navigation: Sequential page numbering added to EN/RU catalog tables
  • Footer cleanup: Custom footer with single title link; duplicate Minima elements removed
  • Badge cleanup: EN/RU shield badges removed from root README files
  • Duplicate H1 fix: Custom layout override + CSS to eliminate duplicate post-header title
  • GitHub Pages compatibility: Explicit pixel dimensions, render_with_liquid: false in _config.yml, absolute SVG URLs

📦 Download

Source code archive attached — unzip and open any .md file in your browser.

🎉 Awesome AI Handbook v1.0.0

Choose a tag to compare

@bestdeejay-design bestdeejay-design released this 24 Jul 18:33

🧠 Awesome AI Handbook v1.0.0

github.com/bestdeejay-design/awesome-ai-handbook — читать онлайн

A practical guide to AI: from running models locally to building your own agents.
Not just links — guides, benchmarks, and proven approaches.

52 topics · ~6 hours of reading · Bilingual (EN + RU)


🚀 Start here

git clone https://github.com/bestdeejay-design/awesome-ai-handbook.git
cd awesome-ai-handbook
# Open README.ru.md or README.md

Or browse online: README.ru.md →


📚 What's inside

Section Files Level Start with
🟢 Zero Level — AI basics, hardware, installation 10 Beginner basics/
🔵 Local Models — inference, selection, tuning 11 Beginner→Advanced local-models/
🤖 AI Agents — architecture, patterns, frameworks 17 Beginner→Expert agents/
💼 Use Cases — coding, RAG, automation, writing 5 Intermediate use-cases/
📖 Resources — communities, courses 5 All levels resources/

✨ Key features

  • 52 detailed guides in Russian (.ru.md) + English (.md) — full translation
  • All code examples work with Ollama + local models
  • Agent tutorials: ReAct agent, CrewAI team, LangGraph coding agent
  • Model catalog: 50+ open-weight LLMs with benchmarks
  • Apple Silicon benchmarks: tok/s for M1–M4
  • Quantization guide: Q4, Q5, Q8 explained
  • Orchestrator deep-dive: OhMyOpenAgent (Sisyphus + 11 agents)
  • Skills & MCP guide: plugins for agents
  • Zero broken links — all cross-references verified
  • MIT License

🏗 Built with

  • OhMyOpenAgent — Sisyphus orchestrator
  • Ollama, Qwen, CrewAI, LangGraph, ChromaDB
  • All content generated and verified locally on Apple Silicon

→ Read the handbook · MIT License