Practical How-To Guides
Step-by-step walkthroughs for AI, automation, and SEO. Every guide ships with the actual commands, payloads, and copy-to-clipboard code you need to use it today.
How Do I Build a Knowledge Graph From My Company's Docs Using Claude?
A knowledge graph lets an LLM answer multi-hop questions across your wiki, CRM, and docs without writing SQL. Here's how to build one from scratch with Neo4j and Claude in an afternoon.
How Do I Build a Multi-Agent Workflow with LangGraph and LangSmith?
Single agents fail at complex tasks. Multi-agent systems work because each agent does one thing well. Here's how to build a research-write-edit team in LangGraph and trace every step in LangSmith.
How Do I Use Claude Code's Slash Commands to Save Hours a Week?
Most Claude Code users miss 90% of the slash commands. Here are the seven I use every day — /init, /compact, /clear, /agents, /hooks, /memory, /resume — and the order to learn them.
How Do I Build AI Agents at Scale with AgentScope?
AgentScope is the open-source framework that takes an LLM agent from notebook to production. Here's how to install it, build a ReAct agent, add tools and memory, and run a multi-agent workflow you can deploy.
How Do I Build a ReAct Agent in Python (Reasoning + Acting Loop)?
Build a ReAct agent in 100 lines, no framework. Once you understand the loop yourself, every agent framework on the market makes sense — and you can debug the ones that don't.
How Do I Build and Ship a Claude Artifact as a Client Deliverable?
Turn a consulting deliverable from a PDF into a working tool. Generate, polish, host, and hand off a Claude Artifact in 30 minutes — and stop shipping docs clients don't read.
How Do I Keep Claude Code Lean in a Large Monorepo?
One 'explain this feature' question shouldn't eat 40 file reads. Configure sub-path CLAUDE.md files, a tight .claudeignore, scoped permissions, and a starting-prompt template to keep Claude Code fast and cheap on any-size repo.
How Do I Run Claude Code on Every Pull Request with GitHub Actions?
Turn every PR into a reviewed PR. A GitHub Action that runs Claude Code against the diff, posts a structured review comment, and can block merges on serious issues. 30-minute setup, weeks of review-cycle time back.
How Do I Keep My Claude Prompt Cache Hit Rate Above 80%?
Prompt caching cuts cached-input tokens to ~10% of normal price — if your prefix is stable. Here's how to structure requests for maximum cache hits, diagnose drift, and stack caching with the Batch API for combined savings.
How Do I Write the Canonical Claude Tool-Calling Loop in Python?
Every Claude agent is a variation of this same 40-line loop. Learn the canonical version, the three gotchas that trip everyone up the first time, and why you should never reconstruct the assistant message by hand.
How Do I Set Up Claude Computer Use Safely on a Mac?
Claude can drive your desktop — or it can drive a sandboxed copy of your desktop. The second option is the only one you should ship. Here's the dedicated-user setup that gives Claude the superpower without giving it your life.
How Do I Decide When to Turn Claude's Extended Thinking On?
Extended thinking produces better answers on hard problems and triples your bill on easy ones. Here's how to classify prompts, A/B test the impact, and route thinking to the 20% of work where it actually earns its keep.
How Do I Give My Claude Agent Persistent Memory Across Sessions?
Make your agent remember what it learned yesterday. A Postgres memory table, read/write/update tools, agent-ID scoping, and a system prompt that writes memories sparingly — not every single turn.
How Do I Build a Research Pipeline with the Anthropic Agent SDK?
A working research agent that searches, fetches, reasons, and writes — end-to-end in Python with the Agent SDK. Structured output, safety rails, and cost tracking from day one.
How Do I Build a Claude-Powered Content Repurposing Pipeline?
Turn one podcast or keynote into LinkedIn, newsletter, blog, and thread drafts — in your voice, overnight. Skill-based pipeline, review pass, voice guide in the loop.
How Do I Schedule Claude Code to Run Maintenance Jobs Overnight?
Bug triage, dep bumps, changelog generation — all perfect for a 3am Claude Code run. Here's the minimum-viable setup: a headless prompt, hard quality gates, and a failure webhook so you know in the morning what happened.
How Do I Set Up Claude Code Hooks That Block Broken Commits Automatically?
Stop catching Claude's half-done work in code review. Hooks fire before edits and commits, run your formatters and tests, and block the action if anything fails. 10 minutes of setup, months of saved review cycles.
How Do I Install and Publish Claude Code Plugins from the Marketplace?
Stop rewriting the same Skills in every project. Claude Code plugins are installable bundles of Skills, MCPs, and hooks. Learn to install one in 10 seconds, then publish your own so the community can use it.
How Do I Connect Claude to My Notion Workspace Without Exposing Everything?
The Notion MCP connector gives Claude live access to your pages and databases. The gotcha: granting the wrong scope shares HR files too. Here's the safe 25-minute setup with read-only defaults.
How Do I Build a TypeScript MCP Server That Exposes My Internal API to Claude?
Your internal admin API already has the tools Claude needs. Wrap it in a TypeScript MCP server in 50 minutes, ship read + write tools with tight input validation, and let Claude use your stack without leaking credentials.
How Do I Package a Team SOP as a Claude Skill Anyone Can Run?
Your team has SOPs nobody follows. Turn one into a Claude Skill and the next person types a slash command instead of hunting through Notion. Consistent output, every time, across every teammate.
How Do I Set Up a Claude MCP Stack for SEO Work?
Connect Claude directly to Google Search Console and Ahrefs via MCP. One prompt pulls live data, reconciles sources, and hands you a prioritized list of quick wins. No CSV exports, no VLOOKUPs.
How Do I Write a CLAUDE.md That Actually Changes Claude's Behavior?
Most CLAUDE.md files are 400-line junk drawers Claude ignores. Here's how to write a short, imperative one that eliminates the corrections you make three times a week.
How Do I Install Claude in Chrome and Schedule a Recurring Browser Task?
Stop manually running the same five-tab research routine every Monday. Install the Claude Chrome extension, set up one scheduled task, and reclaim two hours a week starting next Monday.
How Do I Use Claude's New Memory Feature Without Leaking Client Data?
Claude now remembers across chats by default. For consultants juggling five clients, that's a breach waiting to happen. Here are the two toggles and three habits that lock it down in 15 minutes.
How Do I Cut My Anthropic Bill in Half Using the Batch API?
Anthropic charges 50% of the regular price for batch workloads. Most teams have never turned it on. Here's the one-hour migration that cuts your bill — then how to stack it with prompt caching for deeper savings.
How Do I Automate a Weekly SEO Audit That Emails Me the Wins and Losses?
A cron job that pulls Google Search Console and Ahrefs data, asks Claude for a wins/losses report, and emails it to your inbox every Monday. 45-minute setup, hours saved every week.
How Do I Build an MCP Server That Lets Claude Query My Postgres Database?
Build a read-only Postgres MCP server in an hour. Python, FastMCP, and a hardened database role so Claude can query — but never write, and never blow the budget.
How Do I Build My First Claude Project That Actually Knows My Business?
Stop re-briefing Claude in every chat. In 25 minutes, set up a Project with custom instructions and a knowledge base that locks in your voice and offer — so Claude lands on-brand every time.
How Do I Set Up Claude Code on macOS Without Breaking My Shell Config?
Install Claude Code, Node, and your first CLAUDE.md in 20 minutes without polluting your global shell or sudo-ing your way into permission hell.