Flavio Copes

Learn LangChain and LangGraph with TypeScript: models, prompts, tools, agents, memory, streaming, RAG, graphs, persistence, and human approval.

Cursor Projects gives you a coordinator agent that plans work, delegates it to subagents, keeps shared context, and watches your PRs.

Learn how JSONL stores one JSON value per line, how to write, stream, and query it, and why I use it as Factory Log's event store.

Ghostty from scratch: install it, learn the shortcuts, write the config, pick fonts and themes, use splits, build keybindings, set up shell integration and SSH.

AI made me more productive than I've ever been, but the way building software feels has changed.

ahastack.dev now runs on Astro 7 and htmx 4, has 18 live demos on Cloudflare, and ships a skill file so coding agents know how to build with the AHA stack.

Build native email/password auth on Cloudflare Workers with peppered PBKDF2 hashes, D1 sessions, and hardened HttpOnly cookies.

Learn tmux from scratch: create persistent sessions, use windows and panes, copy text, customize key bindings, and work safely over SSH.

The eight packages in my software catalog are now public MIT repositories on GitHub, with the full source, tests, and documentation in one README each.

CLI Tools is a free, open source macOS app that catalogs every command-line tool you installed with Homebrew, npm, and Cargo, with examples, shell history, and a JSON CLI for coding agents.

An honest report on the Cloudflare products I use in real projects, what each one solves, the gotchas I found, and what I do not use yet.

Typing is slower than speaking, but that friction helps me develop ideas, notice weak sentences, and turn rough thoughts into useful writing.

Things CLI is a free command-line tool for creating, reading, and updating Things 3 projects and tasks from the terminal or a coding agent.

Language Model Builder teaches how LLMs work and lets you train a small GPT locally on an Apple Silicon Mac, without writing code.

MacBooks run hotter above 1800 meters because thin air carries away less heat.

Use Vercel AI Gateway with the AI SDK and OpenAI client, then add model routing, fallbacks, caching, budgets, BYOK, and observability.

Compare 1Password Agentic Autofill, Bitwarden's Agent Access SDK, Browser Use sensitive data, OAuth token vaults, and why passkeys still block AI agents.

exe.dev gives you persistent Linux VMs over SSH, each with an HTTPS URL, private-by-default sharing, and a built-in coding agent.

Learn Zustand through a React store: actions, selectors, immutable updates, persistence, testing, SSR boundaries, and when local or server state is better.

Build and deploy a Content Security Policy with source directives, nonces or hashes, report-only rollout, framing protection, testing, and common CSP mistakes.

Understand passkeys and WebAuthn, then build registration and login with server challenges, stored public keys, SimpleWebAuthn, recovery, and tests.

Learn TanStack Query from the server-state mental model through query keys, freshness, errors, mutations, invalidation, optimistic updates, and common mistakes.

Build a fast documentation site with Astro, Markdown collections, generated navigation, previous and next links, search, and a sitemap.

A practical architecture for AI agents that spend money, deploy code, delete data, or change infrastructure without turning confirmation into a suggestion.

Build and deploy a Workers AI chatbot with the AI binding, validated conversation history, streaming responses, rate limits, cost controls, and state choices.

Follow bun run app.ts through CLI dispatch, module resolution, TypeScript transpilation, JavaScriptCore, Node compatibility, and the event loop.

How Cloudflare Images works: resize, convert and crop images with a URL or a Worker, store uploads, serve private images, and what it costs.

Stream LLM text from a provider through your server to the browser with SSE, a correct buffered parser, cancellation, errors, security, and production checks.

Run an LLM locally with Ollama, choose a model that fits, call the chat API from Node.js, stream replies, request JSON, and understand the limits.

How I use coding agents to understand existing code before changing it: focused questions, tracing data and side effects, and reviews in a fresh session.