AI Lab
Field-tested guides and free tools to build websites and apps, get cited by AI answer engines and work faster. Everything is tested on my own projects, in English, French and Arabic.
Free, no sign-up, and your data never leaves your device.
Title, descriptions and release notes at the right lengths, a check against Google's rules and a check of your graphics before upload.
Open the tool →Create your llms.txt, choose which AI robots may train on, cite or read your site, and analyse your current robots.txt.
Open the tool →Generate an AGENTS.md, CLAUDE.md, GEMINI.md, Cursor rules or Copilot instructions tailored to your stack, with a quality score.
Open the tool →30 professional prompts for development, SEO, marketing, customer support, e-commerce and productivity: fill in the fields, copy, done.
Open the tool →Hours saved, real costs, payback time and a pessimistic scenario: put numbers on an automation before you build it.
Open the tool →Compare the monthly cost of an AI feature across the main providers, with dated, sourced official prices, in dollars, dirhams or euros.
Open the tool →Mask emails, phone numbers, IBANs, cards, ID numbers and API keys before pasting a text into an AI, then put the real values back into its answer. Everything stays on your device.
Open the tool →Which machine can run Llama, Qwen, Gemma or gpt-oss at home? Memory needed, a verdict for each machine and indicative speed.
Open the tool →Describe your business: the tool writes your chatbot's system instructions, a 25-question test plan (including manipulation attempts) and the go-live checklist.
Open the tool →Generate a complete AI policy for your company: approved tools, forbidden data, human review, incidents, with the legal framework of Morocco, France and the EU.
Open the tool →Detailed first-hand accounts: what works, what gets stuck and how to get past it.
The full story of my first Google Play release, Qodex: account, signed AAB, store listing, graphics, declarations, closed test, finding 12 testers and applying for production. The traps I hit, and how to avoid them.
Read the guide →An AI agent can audit, fix and ship dozens of apps for you, or quietly break your project. Here is the method I use every day, with my real wins and the agent's real mistakes.
Read the guide →Answer engines don't replace SEO, they extend it. Here is what is proven, what is likely and what is marketing, with the crawlers to allow, the truth about llms.txt and a 30-day action plan.
Read the guide →Should you pay for an API, take a subscription or run an open model on your own machine? My decision method: privacy, quality, real cost, hardware, security, and the hybrid setup I actually use.
Read the guide →Quote requests, cash-on-delivery orders, email support, content in three languages, invoices: five automations I build for small businesses and online sellers, step by step, with their prompts, safeguards and limits.
Read the guide →Passwords, customer data, documents under NDA: what never goes into an AI assistant, what really happens to your data depending on the plan, and practical methods to use AI without exposing anything.
Read the guide →An overloaded context costs more, answers more slowly and makes more mistakes. Here are the levers I use every day with AI agents, from the simplest to the most advanced, and how to measure what a summary loses.
Read the guide →An AI that only chats is a gadget. Connected to your files, GitHub, your server or your home, it becomes a colleague. Here are the four ways to connect one, how to choose, and how to do it without opening the door to anyone.
Read the guide →The AI guides already published in the OUFRID library.
Tools I build in public, under a free licence. They move fast: their status is stated honestly.
An open format, a command-line tool and an MCP server for AI context files that cite their sources, detect when they go stale and measure what the summary lost.
The Lab is practical, but AI also raises a deeper question: what becomes of the mind when intelligence is no longer a human monopoly? I wrote an essay on it.
Cahiers de la Fracture — Manifeste philosophique en quatre tomes · 2026
An essay on the emergence of a synthetic consciousness and the redefinition of the mind in the age of artificial intelligence. Going beyond the view of language models as mere "stochastic parrots", it puts forward an original conceptual framework: the Oufrid Framework and the Reflexive Closure Criterion. Book in French.
Available on Kobo Apple Books
Coding agents, review, SEO in the age of AI answers.
Build, publish and get an Android app tested.
Save hours without losing control.
Local or cloud, costs, security, vocabulary.
What sets this content apart from "10 amazing AI tools" articles.
Every guide starts from a real project: a published app, a deployed site, a fixed mistake. Failures are told too.
AI attracts hype. I separate what vendors document, what is observed and what remains a guess.
The tools run in your browser. No account, no tracker, no API key to hand over.
Rules change fast (Play Console, AI robots, tools). Every page shows its last update date.
Automations, assistants, multilingual content, apps: I design solutions that are useful, measurable and respectful of your data.