Turn official tutorials into a practical learning path
From Google Gemini and OpenAI GPT to Claude, Hugging Face, and Azure OpenAI. Every guide keeps the official source visible and adds implementation order, safety boundaries, and business exercises.
Local or cloud: record choice conditions without a benchmark
Four explicit fields yield hub starting points or insufficient information. Not a ranking, price or latency test.
Read tutorialOpenAI Agents SDK Quickstart: complete Agent, Runner and inquiry tool
Source includes Agent, @tool and Runner.run. Offline 12-case policy check; openai-agents not installed and no model call.
Read tutorialAgent Skills Quickstart: a valid SKILL.md for synthetic inquiry review
name must equal folder inquiry-review. Offline tests check frontmatter and no-send rules; not installed in Claude Code.
Read tutorialLangChain Quickstart: install with uv, follow official docs
MIT framework. Official README recommends uv add langchain. Don't copy old APIs from stale blogs; LangGraph / LangSmith are related products, not the same package.
Read tutorialOllama Quickstart: run an open model locally
Install the runtime, ollama run a model, chat. API defaults to localhost:11434. Validate on a small model first.
Read tutorialDify knowledge-Q&A tutorial: supported inquiry drafts and unknowns
Six bilingual synthetic FAQs, 26 questions and a complete UI checklist. 22 offline tests passed; Dify and models not run. A citation label is not factual proof.
Read tutorialn8n inquiry-routing tutorial: 12 synthetic cases, review only
Download the workflow and oracle, check inputs, route reviews and handle failures. CLI-tested on n8n 2.37.10; GUI unverified, no customer data or sends.
Read tutorialCursor Quickstart: Tab, Chat and Agent in one IDE
Download the desktop app, sign in, open a project, and ship your first tool with Tab, Cmd+L and Cmd+I.
Read tutorialGemini CLI Quickstart: Google's official terminal agent
Install via npm, sign in with Google, and close one task in your repo. Free tier plus built-in search.
Read tutorialPi Agent vs Claude Code vs Aider: Architecture & Selection Guide
Commercial black box vs open hackability. Comparing tool architecture, SDK embeddability, session trees, and extension cost across three coding agents.
Read tutorialPi Agent Extensions & Skills: Build Your First TypeScript Tool & Workflow
Deep dive into Pi's extension ecosystem: build custom tools with TypeScript, listen to event hooks, and package workflows as Skills.
Read tutorialPi Coding Agent Quickstart: 4 Base Tools & Minimal Terminal Workflow
Break free from commercial black boxes. Master the 4 base tools (read, write, edit, bash) for a clean, transparent terminal workflow.
Read tutorialHermes learning loop: why it knows you better over time
Auto-built skills, self-improvement in use, FTS5 search + LLM summary, Honcho user modeling.
Read tutorialHermes Agent Quickstart: a memory-driven self-improving agent
Nous Research's self-improving agent: full installer, hermes model to switch, learns you over time.
Read tutorialOpenClaw Gateway & channels: Telegram/WhatsApp in one step
The local Gateway is the control plane; channels bring the assistant into your existing chats.
Read tutorialOpenClaw Quickstart: put a personal assistant on your devices
Three steps: onboard, gateway, dashboard. A personal assistant on your own devices.
Read tutorialOpenCode vs Pi vs Claude Code: multi-surface vs minimal vs commercial
Three terminal agents, three routes: open multi-surface, minimal embeddable, commercial managed.
Read tutorialOpenCode Quickstart: the open multi-surface coding agent
Open transparency plus out-of-box polish. Install, connect any LLM, and close one task in the TUI.
Read tutorialMCP in practice: one inquiry tool with list/call shape
The same synthetic inquiry policy as an MCP tool JSON. 12 offline results match the oracle; no live client or SDK transport.
Read tutorialMCP intro: client-server and the three primitives
MCP is the standard protocol connecting AI to tools. Learn the client-server model and three primitives first.
Read tutorialAider vs terminal agents: git flow vs multi-surface vs minimal
Aider = git diff + light; Pi = minimal embeddable; OpenCode = multi-surface; Claude Code = commercial.
Read tutorialAider Quickstart: git-diff pair programming
Every change a reviewable git diff. Install via pip, pick a model, edit by chat.
Read tutorialClaude Code vs open alternatives: commercial or transparent
Closed product, richest ecosystem. Pick managed, or pick auditability with OpenCode/Pi/Aider.
Read tutorialClaude Code Skills workflow: turn team process into SKILL.md
Skills are Claude Code's most valuable asset: write how-to as triggerable, installable files; triggering is ~70% description.
Read tutorialYour First DeepSeek Harness Plugin: Mount It Before You Productize
Follow the official first-plugin tutorial through apply(ctx) and a Cordis overlay. No full step dump — only boundaries and what to do next.
Read tutorialDeepSeek Harness Architecture: Why ‘Everything Is a Plugin’ Is Not a Slogan
Use the official architecture.md to explain profiles, bundles, the session log and turn/step. No package table dump — only where new behavior should attach.
Read tutorialDeepSeek Harness Quickstart: Boot the Web UI, Then Close One Repo Loop
Do not start by writing a plugin. Follow the official README: boot the dsh Web UI, add an API key, pick a workspace, then ask the agent to summarize the repo.
Read tutorialMicrosoft AI for Beginners: Complete 27-Lesson Course (MIT Licensed)
The complete Microsoft open-source AI curriculum: symbolic AI, neural networks, computer vision, NLP, transformers, RL, multi-agent systems and AI ethics — all 27 lessons with links to the original source.
Read tutorialGoogle Gemini API Quickstart: From AI Studio to Your First Request
A practical companion to Google's official quickstart: create a key, protect it, make a text request, and turn the prototype into a maintainable app.
Read tutorialGPT API Quickstart: Build a Reliable Minimum Loop with OpenAI's Guide
Do not start with a complex agent. Secure the key, call the Responses API, handle failures, then add structured outputs and tools one layer at a time.
Read tutorialGPT Prompt Engineering: Turn Good Chat into an Evaluable Workflow
Prompting is not magic phrasing. Use clear tasks, references, constraints, examples, and evals to turn a lucky answer into a repeatable workflow.
Read tutorialClaude API Getting Started: From the First Message to Reliable Long-Context Work
Use Anthropic's official getting-started path to understand the Messages API, system prompts, long inputs, and tool boundaries.
Read tutorialThe Hugging Face LLM Course: From Model Calls to Your Own Evaluation Set
Do not reduce an open-model course to copying a notebook. Build transferable foundations from tokenizers and pipelines to pre-fine-tuning evaluation.
Read tutorialAzure OpenAI Quickstart: Put GPT Behind Enterprise Access and Monitoring
The Azure version is about more than calling a model: resources, deployment names, identity, networking, and logs matter when a prototype becomes an enterprise service.
Read tutorialAnthropic's Prompt Engineering Docs: Claude in Context
Anthropic's official prompt engineering docs are the de facto reference for Claude development: six techniques to scenario templates.
Read tutorialGoogle Gemini Agents Docs: Build Your First Agent
Google's official agent development docs: from concepts to function calling and multi-tool orchestration in the Gemini ecosystem.
Read tutorialOpenAI's Reasoning Models Guide: When & How
Reasoning models (o1/o3) changed prompting rules — no more 'think step by step'. A guide to reasoning effort, structured outputs, and cost control.
Read tutorialGemini API Docs Overview: Models, SDKs & Multimodality
Google AI Studio and the Gemini API docs are one of the best starting points for multimodal LLMs. A guided tour of the docs and a learning path.
Read tutorialClaude 3.5 Sonnet in Practice: From Chat to Agent Workflows
Claude 3.5 Sonnet balanced capability and cost in 2025, letting small teams move agents from experiments to daily tools.
Read tutorialMicrosoft's Generative AI Learning Path
Microsoft's official entry path: from generative AI concepts to Azure practice. A guided tour for building systematic foundations.
Read tutorialHugging Face Agents Course: Official Guide
Hugging Face's official Agents Course — one of the most popular open-source agent tutorials of 2025. Unit-by-unit guide and learning path.
Read tutorialClaude 3.7 Sonnet: Hybrid Reasoning & Coding Agents
Anthropic's first model with both quick answers and extended thinking. A guided reading of the official release and what hybrid reasoning means in production.
Read tutorial