Learn Practical AI. In 10 Minutes a Day.

One short lesson a day. Ten to twenty minutes. Something useful done by the end of each one. Made for people who are new to AI, using ChatGPT and Claude. No coding required.

Boot
LEARNSTACK OS
CHECKING SYSTEM ......... OK
LOADING COURSES ......... READY
LOADING ARTICLES ........ READY
LOADING PROMPTS ......... READY
USER DETECTED ........... WELCOME
READY.
www.learnstackos.comREADY

LIVE PREVIEW

The desktop, in action.

LearnStack OS is a whole desktop built for learning AI. Courses, articles, prompt templates, the tools, and The Hangout each open in their own window you can move and stack, the way a real operating system works.

Watch it click through on its own: open an article, open a course, drop into the chat. All of it is free with an account.

AI in 10 Minutes a Day
10 MIN

One lesson a day, fifteen to twenty minutes. You do the thing yourself on something real, and it is done by the end.

Thirty-eight lessons, seven modules. By lesson 9 you have an assistant that knows your own inbox and calendar, and every lesson after that hands it another job. The first four run on a free account; from lesson 5 you bring a paid ChatGPT or Claude plan.

  1. 4 minLearn One useful AI concept, explained plainly.
  2. 4 minSee it work Follow a practical, familiar example.
  3. 7 minTry it yourself Run it in Claude or ChatGPT on your own email, bill, or question.
  4. DoneDone for today Come back tomorrow, or keep learning now.

One Lesson is enough. For an extra rep, complete the exercise and read today’s new Article.

12Courses34Articles250Prompts119Terms
Latest Articles

Articles on the things that actually change how well AI works for you. Written the same way the Courses are: plainly, and without pretending this is magic.

Article archiveUpdated as new work is published
Courses

12 courses ready for you to dive into. New courses in development.

AI in 10 Minutes a DayA 33-lesson course for people who have never used AI, or tried once and gave up. One lesson a day, ten to twenty minutes, and something useful done by the end of each one. The first 10 lessons run on free ChatGPT and free Claude, and the rest assume a paid plan. You bring a real email, a real bill, a real question, and the work happens on those instead of practice data.AI Proficiency Diagnostic PrepA focused six-part study guide for the AI proficiency assessment: prompt mastery, technical understanding, practical application, critical evaluation, workflow design, and systems building. Each part drills the concepts and the judgment the assessment rewards, with practice built in.AI Foundations: Zero to AI HeroStart with no AI experience and build practical proficiency across ChatGPT/Codex and Claude/Claude Code. Learn how prompts, context, verification, personalization, multimodal work, connected knowledge, and AI agents work, then turn successful experiments into repeatable systems. Every lesson produces a practical deliverable, taking you from your first useful conversation to a portable personal AI operating system.Getting Started with CodexSet up Codex safely, then practice seven starter workflows with clear boundaries and proof at every step. You go from a careful first install to project rules, packaged skills, run specs, and a real artifact you built. The final lesson audits your first week so you know exactly what to trust it with next.AI for Small Business & SolopreneursPractical AI for owner-operators who wear every hat: customer replies, marketing copy, quotes, hiring posts, and simple automation. You learn the prompt recipe once, then apply it to the jobs that eat your week. It ends with reusable workflows and a clear-eyed lesson on when not to trust the output.ChatGPT & Claude Power UserGo past casual chatting and turn ChatGPT and Claude into a configured working environment. You set up custom instructions and memory, pick the right model for each job, work inside projects with your own files, and chain multi-step workflows. It closes with the trust boundary: what to verify before you rely on it.AI for Real Estate AgentsAI for working agents: listings that sell, a week of marketing content in one sitting, lead follow-up that never drops, and comps translated into plain English for clients. Compliance is built in, with a dedicated lesson on fair housing and a system for staying inside the lines. It ends with a weekly prospecting engine you actually run.AI for Real Estate InvestorsAI for deal-driven investors: screen opportunities against your buy box in minutes, research markets with verification built in, and triage leases fast. You learn to draft offers and counter-offers with full deal context loaded, while never trusting the math blind. It ends with your entire portfolio summarized on one page.AI for Content CreatorsPractical AI for independent creators: ideate against the blank page, draft from your own rough notes, and turn one strong piece into a week of content. The voice card lesson keeps everything sounding like you, not like a machine. It closes with the audience-trust line you should never let AI cross.AI for Teachers and EducationPractical AI for teachers and tutors: plan lessons, differentiate one worksheet into three levels, and draft rubrics, feedback, and parent messages in your own voice. Student privacy is protected by design, with clear rules on what never goes into a prompt. It ends with a weekly prep routine that gives you your evenings back.AI for Job SeekersPractical AI for the job hunt, honest by design: build your professional story from real experience, tailor resumes and cover letters to each posting, and prep for interviews with targeted practice. Nothing is invented — the machine reshapes your true material instead of writing fiction. It ends with a weekly search routine that keeps momentum without burnout.AI for NonprofitsPractical AI for lean nonprofit teams: draft grant narratives, donor appeals, and board reports from your own real data, then run communications in a fraction of the time. Nothing is ever fabricated — impact numbers come from your records or they do not appear. It ends with reusable workflows a two-person team can actually sustain.
Prompt Library

Copy-ready templates for real work — browse the library.

Contract SummarizerProduces a structured, flag-first summary of any contract: parties, money, deadlines, obligations, and nonstandard terms.Direct-to-Owner Outreach LetterDrafts a letter to a property owner that sounds like a person wrote it, states an honest premise, and makes no reply acceptable.Debt Payoff Strategy ComparisonCompares payoff orderings for multiple debts with the actual arithmetic, so you choose a strategy on numbers and psychology, not folklore.Elevator Pitch BuilderBuilds a clear, memorable elevator pitch that says what you do and why it matters in the time you actually have.Voice MatcherTeaches the model your writing voice from two real samples, then drafts new content that sounds like you.Engagement Status ReportWrites the weekly engagement status that keeps sponsors informed, risks visible, and surprises impossible.
Skills Library

Skill files that teach your AI assistant a repeatable job — sign up to use them.

Prompt Library

A growing library of 250+ practical prompts

Every template here was built for real work: real estate deals, small business ops, legal reviews, marketing, finance, and a dozen fields between. Each one ships with notes on why it's built that way and what to watch for. Copy one, paste it into the AI you already use, and swap the bracketed placeholders for your own numbers, clients, and cases.

Contract Summarizer1 / 4
Summarize the contract between the tags. Use exactly this structure:

1. PARTIES AND ROLES: who is who, one line each.
2. MONEY: every amount, who pays whom, when.
3. DEADLINES: every date and time-triggered obligation, as a table with columns Date/Trigger, Obligation, Consequence if missed.
4. MY OBLIGATIONS: everything {MY_ROLE} must do.
5. THEIR OBLIGATIONS: everything the other party must do.
6. EXIT: how each party can terminate, and what it costs.
7. FLAGS: anything unusual, one-sided, or missing that a careful reader would question. Mark each [FLAG].

Rules: base everything only on the document. If something standard is absent (e.g. no termination clause), flag the absence. Quote section numbers for every flag. Do not soften.

<contract>
{CONTRACT}
</contract>

NOTES — Section 7 is why this exists: models are good at spotting one-sided terms when explicitly told not to soften. This is a reading aid, not legal review; the flags are your agenda for the attorney, not a substitute for one.

LearnStack Apps

From a rough idea to a clear prompt or a finished image. Find both apps in your LearnStack Toolbox.

Illustration of a sunlit prompt workstation turning a short idea into a structured document.

Prompt Generator

Turn your rough idea into a detailed prompt with a clear role, objective, and output format. Copy the result into your preferred AI tool and make it your own.

Illustration of a sunlit image workstation creating a colorful landscape and a matching print.

Image Generator

A Premium tool for turning your ideas into images. Describe what you have in mind, choose a style and aspect ratio, then preview and download the result within your available allowance.

AI Dictionary
AI Dictionary

119 AI terms defined in plain English, cross-linked from every lesson.

Acceptable Use PolicyThe provider's rules for what you may and may not do with their AI service; violations risk account suspension.AgentAn AI system that pursues a goal by taking multiple actions on its own: planning steps, using tools, checking results, and continuing until done.Agentic WorkflowA multi-step process where AI carries out a sequence of decisions and actions with limited human involvement, often across multiple tools.AI DetectionTools that claim to identify whether text or media was AI-generated; for text, their reliability is poor.AI-Generated Content DisclosureTelling your audience when content was produced by AI, whether required by a platform, a law, or plain trust-keeping.APIThe programmatic doorway to a model: software sends a prompt to the provider's servers and gets the response back, no chat window involved.API KeyThe secret credential that identifies your account on API calls; whoever holds it can spend your money.Artificial Intelligence (AI)Software that performs tasks which normally require human thinking, like understanding language, recognizing patterns, or making decisions.BenchmarkA standardized test used to score and compare AI models on tasks like math, coding, or reading comprehension.Chain-of-Thought PromptingAsking the model to show its reasoning step by step before giving a final answer, which measurably improves accuracy on hard problems.Chatbot vs. API AccessThe two ways to use the same model: a chat product with a flat subscription and built-in features, or raw API access billed per token with full control.ChunkingSplitting long documents into smaller pieces before embedding them, so retrieval can find and return the precise relevant passage.CitationA pointer from a model's claim back to the source that supports it, letting you verify instead of trust.Code Execution SandboxA sealed environment where an AI can write and actually run code (calculations, data analysis, file processing) without touching your systems.Computer Use (Browser Automation)An AI capability where the model operates software like a person would: viewing the screen, clicking, typing, and navigating websites to complete tasks.ConnectorA packaged integration that links an AI assistant to a specific external service, like your email, calendar, or file storage.ConstraintA hard boundary set in the prompt: length limits, banned words, required elements, scope restrictions.Content ModerationAutomated screening of prompts and outputs for policy-violating content, applied by providers and available to builders.ContextAll the background information you give the model so it can do the task: documents, facts, history, and constraints.Context CachingA cost-saving feature where a provider stores a repeated portion of your prompt so subsequent calls reusing it are billed at a steep discount.Context TruncationWhat happens when input exceeds the context window: the overflow is silently cut off, usually from the oldest material.Context WindowThe maximum amount of text, measured in tokens, that a model can consider at once: your prompt, documents, conversation history, and its reply combined.Custom InstructionsStanding preferences you write once in a chat product's settings, applied automatically to every conversation.Data Privacy (Training on Your Data)The question of what an AI provider does with what you send it: whether your prompts are stored, reviewed, or used to train future models.Data RetentionHow long a provider stores your prompts, files, and outputs after processing, and what can happen to them during that window.DelimiterMarkers like triple quotes, dashes, or tags used in a prompt to clearly separate instructions from content.EmbeddingA list of numbers representing the meaning of a piece of text, so that similar meanings produce nearby numbers.EndpointThe specific web address a piece of software calls to use a particular API function, like sending a chat message or creating an embedding.ExemplarA single example of ideal output included in a prompt for the model to imitate.Extended ThinkingA mode where a model spends additional visible or budgeted effort reasoning through a problem before producing its final answer.Few-Shot PromptingIncluding several input-output examples in your prompt so the model learns the exact pattern you want.Fine-TuningAdditional training on a focused dataset to specialize a general model for a particular task, style, or behavior.Foundation ModelA large general-purpose model that serves as the base for many different applications and specialized versions.Frequency PenaltyA setting that discourages the model from repeating words it has already used, scaling with how often each word has appeared.Frontier ModelOne of the most capable AI models available at a given time, sitting at the leading edge of the field.GroundingAnchoring a model's answers in supplied source material (documents, data, search results) instead of its trained memory.GuardrailsThe layered safety measures that constrain what an AI system will say and do: trained-in values, filters, and application rules.HallucinationWhen a model states something false with full confidence: invented facts, fake citations, made-up numbers.Human in the LoopDesigning AI workflows so a person reviews or approves the consequential steps instead of letting automation run unchecked.Image GenerationAI creating new images from a text description, or editing existing images according to instructions.
AI News
NVIDIA DeveloperSEP 18Benchmarking LLM Inference at Scale with AIPerfYou’re deploying a model on a system. It starts up, prompts are getting responses. Now the hard question: Is this fast? Your instincts might lead…Google AISEP 18New experts join Google’s AI & Economy teamimg src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/AI__Economy_team_hero.max-600x600.format-webp.webp" We are expanding our AI…Google AISEP 18Co-creating the future of fashion with Googleimg src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/Blog_Header_V2.max-600x600.format-webp.webp" Google worked side-by-side…TLDR AISEP 18Claude Projects v2 💼, Google family agent 👨‍👩‍👧‍👦,Google AISEP 17Making global data easier to exploreimg src="https://storage.googleapis.com/gweb-uniblog-publish-prod/images/data-commons.max-600x600.format-webp.webp" Google and the UN system have…OpenAISEP 17How Cooley is accelerating IPO work with ChatGPTCooley built GO Public with ChatGPT Work to bring intelligence to the IPO process, helping lawyers surface issues earlier and focus judgment where…OpenAISEP 17Introducing Astra for LawOpenAI for Law brings frontier intelligence for law, custom firm workflows, connected legal data sources, and legal-grade controls for confidential…TLDR AISEP 17Claude + Cowork merge 🛠️, ChatGPT sponsored agents 💰, harness tax 🤖NVIDIA DeveloperSEP 16How to Use AI Agents to Prepare 3D Scenes for SimulationAgentic AI workflows can be used to prepare and validate digital twins for physical AI systems. Agents can inspect 3D scenes, author…NVIDIA DeveloperSEP 16TensorRT Edge-LLM Completes the MLPerf Edge Agentic Benchmark 6.4x Faster on Jetson AGX ThorAI agents are moving from cloud data centers to vehicles, robots, and other edge devices. Unlike a chatbot that answers a single prompt, an agent…
AI News
select a headline to open it in Articlespauses while you explore
The Hangout
The Hangout — sample conversation
  • #general-chat
  • #daily-boost-help
  • #prompt-sharing
  1. renee_opsAnyone have a prompt for summarizing inspection reports?
  2. marcus.kYes — posting mine in #prompt-sharing. It's the Lesson 7 template.
  3. LearnBotHappy to help if you get stuck. I'll be here. Coffee's fresh.
  4. dana.pNew here — where should I start if I've never used AI at work?

The Hangout

Where the operators talk shop.

Courses teach you the system. The Hangout is where you compare notes on running it. Members swap prompts that actually worked, ask the questions tutorials skip, and find the right place to ask when you get stuck. LearnBot hangs out here too, mostly drinking coffee.

Start Chatting

MEMBERSHIP

Join for free. Optional premium upgrades available.

FREE ACCOUNT

Start learning for free with a free account. Courses, articles, prompt templates, the tools, The Hangout, and progress that saves between sessions. No card, nothing to cancel.

Free Courses and Lessons, in short daily steps

The full library of copy-ready prompt templates

Toolbox Skills, DIY Projects (Coming Soon), and the AI Dictionary

The Hangout community, plus notes and progress that save

README
Bhaskar Pandey, founder of LearnStack

I built LearnStack because too much AI training asks for hours you do not have or races straight to breathless predictions about the future of work.

Neither one helps. What helps is ten minutes a day, in order, in plain English, with tools you can use the same afternoon.

So that's what this is. No hype about the future of work. No pretending these tools are smarter than they are, because knowing exactly where they fail is most of the skill. Just the practical part, arranged so you'll actually get through it.

Bhaskar Pandey, Founder, LearnStack

READY.

Ten minutes. Right now. No account.

The first Lesson is open. If it's not for you, you've lost ten minutes and gained a prompt template.

A growing library of articles, prompts, and skills

DIY projects are Coming Soon

Prompt Generator to convert your one-liners to detailed prompts

AI News from the top sources in one place

Live Webinars on latest topics in AI

Masterclasses on vibe coding, building automation agents, and more

LearnBot guides every lesson. One lesson a day, fifteen to twenty minutes, plain English, no prerequisites.

Start the first Lesson