A real AI tutor, not another chatbot. Give it a goal and Ignia builds you a structured course, then teaches you through it, reshaping every session around what you have actually learned.
Builds your course Adapts as you go Free and open source
What makes Ignia different
Most AI learning tools answer your questions on demand. Useful, but no different from a search engine with a personality. Ignia works the other way around: it builds a structured course around your goal, teaches you through it, and keeps adapting as you go.
Chat AI (ChatGPT, Claude, ...)
Answers one question at a time. No course, no memory of what you know, no plan.
Ignia
Builds a full course from your goal. Tracks what you know, schedules what to review, and reshapes the next session around where you struggled.
Flashcard apps (Anki et al.)
Right or wrong, repeat. Good for memorizing. Stops there.
Ignia
Pushes past recall. You explain, apply, analyze, find the bug. When you write a free-text answer, the AI reads it and tells you exactly what to improve.
Language apps (Duolingo et al.)
Built around streaks, points, and daily habits. Learning is the side effect.
Ignia
Built around actual learning. Short sessions are fine when the review schedule is right. No streaks, no leaderboards.
Tell Ignia your goal, what you already know, and what you have on hand: a book, a video, notes, anything. It figures out the rest.
Ignia writes a structured plan with clear learning objectives. Take a short test to skip what you already know. Edit the plan freely; it is yours.
Each session uses exercises matched to what you are working on, from simple recall to real application. The app tracks how you do, brings the hard pieces back at the right moment, and reshapes the next session around your gaps.
Flashcards train recall: that is the floor, not the ceiling. Real understanding means being able to explain, apply, and spot errors. Ignia matches the exercise to the concept and to the level you are working at. The AI adapts the mix as you go.
Get the fact back from memory. The starting point, not the finish line.
Flashcard
A question on one side, the answer on the other. Rate how well you knew it; the app schedules the next review.
Fill the blank
"The French Revolution began in ___, with the storming of the Bastille." One word or phrase to complete the sentence.
Multiple choice
Pick the right answer among plausible options. Used sparingly: easy to guess without actually knowing.
True or false
A statement about the course material. Agree or disagree. Designed around the misconceptions the subject actually creates.
Show you have grasped the structure, not just the surface.
Matching
Pair items from two lists: a term with its definition, a date with its event, a word with its translation.
Categorize
Sort a set of items into the right groups: types of sentences, chemical families, kinds of government.
Ordering
Put a sequence in the right order: steps in a process, events in history, stages of an argument.
Word order
Rearrange scrambled words into a correct sentence. For languages, this tests grammar, not just vocabulary.
Use what you know to produce an answer, not just recognize one.
Short answer
A question that needs a real answer in your own words: one sentence or a few. Graded automatically.
Numeric answer
Calculate a result: a speed, a price, a measurement. Include units where needed.
Equation solving
Solve a math or physics problem step by step. The app checks whether your result is equivalent, not just identical.
Code trace
Read a short piece of code and predict what it will output. Tests whether you actually understand what the code does.
The deepest level. Produce something, or diagnose what is wrong.
Written response
AI feedbackExplain a concept, make an argument, describe a process in several sentences. The AI reads your answer and tells you specifically what to improve.
Code review
AI feedbackFind what is wrong in a piece of code and explain how to fix it. The AI checks both the diagnosis and the proposed solution.
The types keep growing as new formats are validated. The point is that the right exercise gets used for the right concept at the right level. Not a uniform deck for everything.
Ignia needs an AI at one point: to write your course and prepare each session. That can be a hosted service, your own API key, or a model on your machine. After that, the daily work runs offline, and on paper if you want.
No connection needed
Once a session is generated, you practice with no connection at all. The app grades most exercise types on its own: matching, ordering, fill the blank, equations, code traces. Export any course to read, print, or work through on paper.
Where the AI comes in
Writing the course and preparing sessions is the one step that needs an AI. It keeps adapting too: after each session it sees where you struggled and rebuilds the next one around your real gaps. You can talk to it: ask for a different explanation, edit the plan, or go deeper on a topic you found hard.
The science behind the sessions
Repetition alone does not make a learner. Learning science has identified specific techniques that produce durable knowledge, and they are often the opposite of what feels natural or rewarding in the moment. Ignia applies them systematically, every session.
The app knows when you are about to forget something and brings it back just in time. Not on a fixed weekly schedule, but based on how well you actually know it.
Cepeda et al., 2006; Roediger & Karpicke, 2006
Sessions mix different topics and concepts rather than finishing one before starting the next. It feels harder, and that is the point. Mixing is what makes knowledge actually stick.
Rohrer & Taylor, 2007; Bjork & Bjork, 2011
The app tracks your success rate and adjusts. Too easy and exercises get harder. Struggling and it slows down, surfaces the prerequisites. The target is the zone where learning actually happens.
Bjork, 1994
Topics break down by what kind of thinking they require: remembering a fact, understanding why, applying a method, analyzing something new. Each concept gets exercises at the right level, not a one-size-fits-all quiz.
Anderson & Krathwohl, 2001
A wrong answer comes back with what was off, why, and what to look at next. Not just a red mark. If you think the AI got it wrong, you can explain your reasoning and it reconsiders.
Hattie & Timperley, 2007
No streaks. No leaderboards. No "mastered in 30 days." Points and badges have real limits as motivators, and they say little about what you have actually learned. What Ignia shows instead is your honest retention, which dips when you stop practicing.
Dichev & Dicheva, 2017
Everything Ignia creates is yours to keep, in a form you control. Export any course as a plain document to read, print, or share. Reach for a top AI service when you want the best content, or run a local model when you would rather keep everything on your machine. Privacy here is not an upsell; it is simply how the app is built.
Everything Ignia creates, your plan, course, exercises, and progress, is stored on your computer. No account needed to get started.
Use any major AI service for the best content quality, or run a model entirely on your own machine for full privacy. You decide per course, not per account.
If you prefer, install Ignia on a home server or shared machine and reach it from anywhere. One command to set up; nothing goes through our infrastructure. This is optional, not the default path.
Export any course as a clean document. Your course sections and exercises in a single file you can read in bed, print, or annotate. No app needed.
Export a full course as a shareable file. Give it to a friend; they open it in their own Ignia and start practicing immediately, on their own machine.
The app is fully open source. Read what it does, verify it, contribute to it. No lock-in. The project is the product.
Ignia is open source. The code, the pedagogy choices, and the technical bits are all in the repo. The docs explain how it fits together.