Skip to main content

Command Palette

Search for a command to run...

Why I Built "Learning Space AI": Organizing the Chaos of LLM Chats

Stop doom-scrolling. Start building a curriculum. A local-first workspace for intentional learning.

Updated
4 min read
A

I’m a visually impaired software engineer who finds deep joy in exploring ideas and uncovering unexpected connections. I’m drawn to patterns that often go unnoticed. I love finding those veered threads that do not seem related until they suddenly come together. I write to make sense of what I learn and enjoy breaking things down for others. For me, it is about connecting ideas, sharing the process, and letting curiosity lead the way.

If you’ve ever tried to learn something seriously using ChatGPT, you probably recognize this pattern: You open a tab. Then another. Then another. Each conversation is useful, but disconnected. Promising threads get abandoned. Good explanations get buried. After a while, learning starts to feel busy instead of cumulative. I built Learning Space AI to solve that exact problem. It’s not another chatbot. It’s a way to plan, organize, and complete learning with intention—while still using ChatGPT exactly where it works best.

What Learning Space AI Is (and Isn’t)

Learning Space AI helps you separate thinking about learning from talking to AI. It gives you a dedicated workspace to:

  • Decide what you want to learn.
  • Design good prompts ahead of time.
  • Execute them one by one in ChatGPT.
  • Capture what you actually learned. It does not:
  • Auto-send prompts in the background.
  • Replace your thinking with automation.
  • Lock your data in a proprietary cloud.

    The Core Idea: Containers for Knowledge

    The app is built around two simple concepts:

    1. Projects (The Context)

    Projects are containers. They might represent a course you’re following, a book you’re reading, or a goal like "Learn Rust Basics." The Secret Sauce: You can set a System Prompt for each project (e.g., "You are a Systems Engineer. Prioritize memory safety."). You set this once, and the app applies it to every interaction in that project automatically.

    2. Learning Cards (The Question)

    Learning cards are individual tasks. Each card contains a specific prompt (e.g., "Explain the Borrow Checker"). One card = one question = one learning intention. This prevents the "everything in one conversation" problem and encourages focus.

    The Workflow: Calm & Deliberate

    We treat manual thinking and AI assistance as equal, first-class options.

    1. Create Cards: Manual or AI

    You can write cards yourself if you know what you want to ask. But if you don't know where to start, click "🤖 Generate".
  • The app opens ChatGPT with a curriculum-design prompt.
  • You talk to the AI to design a study plan.
  • You copy the JSON code block it produces and paste it directly into the app.
  • Result: An instant study plan, curated by you.

    2. Launch: Learning happens in the browser

    When you’re ready to learn, you click Start Chat (▶️) on a card.
  • Your default browser opens to ChatGPT.
  • The prompt is pre-filled (combining your Project Context + Card Question).
  • You press "Send" yourself. This small bit of friction is intentional. It ensures you always know what is being sent and stay in control. Learning remains an active choice, not a background process.

    3. Capture: Notes turn conversations into understanding

    A good AI explanation doesn’t count as learning unless you process it. Don't just copy-paste. Switch back to the app and click Open Notes. The app features a rich Markdown editor where you can:
  • Summarize the answer in your own words.
  • Save code snippets using code blocks.
  • Format key points with bold and italics.

    4. Complete: Visible Progress

    Once you understand the concept, mark the card as Done. It moves to the Completed section. This creates something many AI learning workflows lack: a visible sense of progress. You aren't just chatting; you are finishing things.

    Who is this for?

    Learning Space AI is for people who:
  • Learn intentionally.
  • Use ChatGPT as a thinking partner, not a shortcut.
  • Prefer structure over chaos.
  • Want learning artifacts (notes), not just chat history. It is likely not for you if you want fully automated learning or "set it and forget it" AI agents.

    Under the Hood: Svelte 5 & Electron

    For the developers reading this, this project is built with Electron, Svelte 5, and TypeScript.
  • Svelte 5 Runes: We use the new $state and $derived syntax for clean, reactive state management.
  • Local-First: Data is stored in simple JSON files (LowDB) on your machine.
  • Accessibility: The app is built from the ground up to be fully keyboard and screen-reader accessible.

    Try It Out

    If you are tired of losing track of your AI chats, give it a shot. It’s free, open-source, and respects your privacy.
  • 📦 Download the Installer: GitHub Releases
  • ⭐ Star the Repo: github.com/adil-adysh/learning-space-ai Learning works best when it’s intentional. Learning Space AI exists to protect that intention.
Why I Built "Learning Space AI": Organizing the Chaos of LLM Chats