How to Remember What You Read: A Science-Based Guide
Talk this through with a Chapterly AI tutor
Ask anything about this article. Your first 2 messages are free — no signup required.
Quick Answer: To remember what you read, use three science-backed strategies: spaced repetition (reviewing highlights at increasing intervals right before you'd forget them), active recall (testing yourself before re-reading), and elaborative interrogation (asking "why?" and "how?" as you read). Adding just 15–20 minutes of processing per book—a brief chapter checkpoint plus daily highlight review—can increase long-term retention from 10% to 70%+.
You've just finished an incredible book. The ideas were transformative, the insights profound. You tell yourself you'll never forget these lessons.
Fast forward a month. Someone asks you about the book. You remember enjoying it, but the specific ideas? They've faded into a vague blur.
This isn't a personal failing—it's how human memory works. Research shows we forget approximately 70% of new information within 24 hours and up to 90% within a week. This phenomenon, known as the "forgetting curve," was first documented by psychologist Hermann Ebbinghaus in 1885, and it affects everyone.
But here's the good news: the same research that identified the forgetting curve also revealed how to beat it.
The Science of Forgetting (And Remembering)
The forgetting curve isn't random. It follows a predictable pattern, which means we can strategically intervene to flatten it. The key insight from learning science is that memory isn't about passive absorption—it's about active engagement.
Every specific number and named study in this guide is real and citable — Ebbinghaus's original 1885 forgetting-curve experiments, the 2015 Murre and Dros replication, Nesbit and Adesope's 2006 concept-mapping meta-analysis, Rayner et al.'s 2016 speed-reading review, and others named at first mention — so you can trace any claim back to its source rather than take our summary of it on faith.
Why We Forget
When you read something once, your brain creates a weak neural pathway. Without reinforcement, this pathway degrades rapidly. It's like walking through tall grass: the first pass barely leaves a trace, but repeated walks create a lasting path.
Several factors accelerate forgetting:
- Passive reading: Simply moving your eyes across words without engaging deeply — including highlighting without retrieval practice
- No review: Never revisiting the material after the initial reading
- Lack of connection: Not linking new ideas to existing knowledge
- Context switching: Jumping to new content before consolidating what you just learned
How Memory Actually Works
Memory consolidation happens through a process called "retrieval practice." Every time you actively recall information (rather than just re-reading it), you strengthen the neural pathway. This is why testing yourself is far more effective than reviewing your notes.
The spacing effect compounds this benefit. When you spread out your review sessions over increasing intervals, each retrieval strengthens the memory more than if you crammed all your review into one session.
Strategy 1: Spaced Repetition
Spaced repetition is the most powerful technique for long-term retention. For a deeper dive into this method, see our complete guide to spaced repetition for readers. Instead of reviewing material all at once, you review it at gradually increasing intervals: first after one day, then three days, then a week, then a month.
How It Works
The optimal time to review something is right before you're about to forget it. Each successful retrieval pushes the memory further into long-term storage and extends the interval until the next review is needed.
Here's a practical schedule for book highlights:
- Day 1: Review your highlights from today's reading
- Day 3: Review again
- Day 7: Review again
- Day 14: Review again
- Day 30: Review again
- Day 60+: Review at increasingly longer intervals
Making It Practical
The challenge with spaced repetition is tracking what needs review and when. This is exactly why we built Chapterly—to automate this process for readers. The app schedules your highlight reviews at optimal intervals, so you never have to think about timing.
If you're doing it manually, consider:
- Using a simple spreadsheet to track review dates
- Setting calendar reminders for review sessions
- Creating a physical review stack with dated dividers (the Leitner system for readers formalizes this with five boxes and clear advance/reset rules)
A 2025 classroom study puts a number on how much this discipline is worth. Dobson (Advances in Physiology Education, 2025) ran two sections of the same university course side by side, identical except that one section received a short tutorial on retrieval and distributed practice plus regular low-stakes pop quizzes, while the other did not. The instructed section scored significantly higher on exams, and quiz performance correlated strongly with exam performance (r = 0.63). The intervention itself was not exotic — a 30-minute explanation of why spacing and self-testing work, plus quizzes that forced students to actually do it. That is a reasonable description of what this entire guide is asking you to build for your own reading.
Strategy 2: Active Recall
Active recall means testing yourself rather than passively re-reading—it's one of several proven active reading strategies that transform retention. When you try to retrieve information from memory before looking at it, you strengthen the memory far more than if you simply reviewed it.
The Testing Effect
Studies consistently show that students who test themselves remember significantly more than those who re-read material—even when the testers spend less total time studying. This "testing effect" is one of the most robust findings in learning science. A 2021 meta-analysis by Yang and colleagues, published in Psychological Bulletin, pooled 222 studies and nearly 48,500 students and found that classroom quizzing raised academic performance by a medium-to-large margin (g = 0.499) across 18 subject areas and every education level from elementary school through college—evidence that the effect holds at real-world scale, not just in small lab studies.
How to Apply It
After finishing a chapter:
- Close the book and write down the main ideas from memory
- Try to explain the key concepts as if teaching someone else
- Ask yourself questions: What was surprising? What contradicted what I already believed? How does this connect to other things I know?
For your saved highlights:
- Cover the highlight and try to recall what it says
- Quiz yourself on the key concepts before looking at your notes
- Use the Feynman Technique: explain ideas in simple terms — or go further with the protégé effect and read with the intent to teach, which research shows improves retention by up to 28%
Strategy 3: Elaborative Interrogation
Simply asking "why?" and "how?" as you read dramatically improves retention. This technique, called elaborative interrogation, forces you to connect new information to existing knowledge. It is closely related to the self-explanation effect — the research-backed finding that explaining material to yourself as you read produces 20–40% better retention than passive reading.
Why It Works
When you explain why something is true or how it connects to other ideas, you create multiple retrieval pathways to the same information. Instead of a single fragile connection, you build a web of associations that makes the memory more durable.
Practical Applications
As you read, pause regularly to ask:
- Why does this make sense?
- How does this connect to something I already know?
- What would be an example of this in my own life?
- How does this contradict or support other ideas I've encountered?
This is where AI reading assistants can be particularly valuable. Having a discussion partner after each chapter helps you process and elaborate on the material in ways that are difficult to do alone.
The technique compounds fastest on aphoristic or argument-dense texts, where nearly every passage is doing real cognitive work rather than advancing a plot. Marcus Aurelius's Meditations is a useful test case, since the book is itself a record of its author asking himself these same "why" and "how" questions: readers who pause on each entry and interrogate the specific fear or attachment it is arguing against consistently retain more than readers who move through it front-to-back like a devotional. (Our Meditations summary and study guide groups the entries by recurring theme, which makes this kind of interrogation easier to sustain across the whole book.)
Strategy 4: The Chapter Checkpoint
Create a ritual at the end of each chapter where you stop and process what you just read. This checkpoint prevents the common problem of "finishing" a book while actually remembering very little.
The Checkpoint Process
- Summarize: Write 2-3 sentences capturing the main point — paraphrasing in your own words rather than reusing the author's phrases is the most reliable version of this step
- Extract: Identify the most important highlight or quote
- Connect: Note one connection to something else you know
- Question: Write one question the chapter raised for you
This takes about 5 minutes but dramatically increases retention compared to immediately moving to the next chapter.
Strategy 5: Interleaving
Instead of binge-reading one book, interleave multiple books on related topics. This technique forces your brain to compare and contrast ideas, creating deeper understanding.
How to Apply It
- Read 2-3 books on the same topic simultaneously
- Alternate between them rather than finishing one before starting another
- Actively look for connections and contradictions between authors
- Let the ideas "cross-pollinate" in your mind
Strategy 6: The Feynman Technique for Readers
Named after Nobel Prize-winning physicist Richard Feynman, this technique is one of the most powerful ways to ensure you truly understand what you read—not just recognize it. For a deep dive, see our complete guide to the Feynman Technique for reading.
The Four Steps
- Choose a concept from your reading—an argument, framework, or key idea
- Explain it in simple language, as if teaching a 12-year-old. No jargon, no hiding behind complex vocabulary
- Identify gaps where your explanation breaks down or gets vague. These gaps reveal what you don't actually understand
- Return to the source and fill those gaps, then simplify your explanation further
Why It Works for Books
Most readers confuse recognition with understanding. You recognize a concept when you see it again, but recognition fades quickly. The Feynman Technique forces generation—you must construct the explanation from scratch, which creates far deeper memory traces.
This is also why AI reading assistants can be valuable. Having a discussion partner that asks follow-up questions naturally implements the Feynman Technique—you're forced to explain, defend, and refine your understanding of what you just read.
Strategy 7: Note-Taking Methods That Aid Memory
Not all note-taking methods are created equal. Research shows that how you take notes matters more than how much you write. The right method also depends on what you are trying to keep: retaining a book's argument is a different problem from remembering the names, dates, and facts it is built on, which rewards more targeted techniques.
The Cornell Method for Books
Divide your note page into three sections:
- Right column (70%): Main notes while reading—key ideas, arguments, evidence
- Left column (30%): Keywords, questions, and cues you write after reading
- Bottom section: A 2-3 sentence summary of the chapter
During review, cover the right column and use the left column cues to practice recall. This builds active recall directly into your note-taking.
Concept Mapping
For books with complex, interconnected ideas, draw concept maps showing relationships between ideas. Research by Nesbit and Adesope (2006) found that concept mapping outperformed note-taking, re-reading, and attending lectures for long-term retention of complex material.
The Zettelkasten Method
Originally developed by German sociologist Niklas Luhmann, this method involves creating atomic notes (one idea per note) with links between related ideas. Over time, this builds a personal knowledge network that mirrors how your brain actually stores information. See our Zettelkasten for readers guide for a practical introduction, or our guide on turning highlights into smart notes for a workflow-focused approach.
How Much Do We Actually Forget?
The research on forgetting is both sobering and motivating. Murre and Dros (2015) replicated Ebbinghaus's original forgetting curve experiments with modern methods and confirmed the findings still hold:
- After 20 minutes: ~42% forgotten
- After 1 hour: ~56% forgotten
- After 1 day: ~66% forgotten
- After 1 week: ~77% forgotten
- After 1 month: ~79% forgotten
However—and this is the crucial part—each review session dramatically flattens the curve. After just 3-4 well-timed reviews, retention jumps to 80-90% even after months. The strategies in this guide are designed to trigger exactly those well-timed reviews.
A Worked Example: One Book From Start to Long-Term Memory
Abstract strategies are easier to apply once you see them chained together on an actual book. Here is what the system above looks like end to end, using James Clear's Atomic Habits as the example — chosen because most readers finish it feeling like they understood it, then cannot reproduce the actual framework six months later. (If you have already read it, books like Atomic Habits covers where to go next.)
While reading Chapter 3 ("How to Build a New Habit in 4 Steps"): You highlight the cue-craving-response-reward loop and the four laws (make it obvious, make it attractive, make it easy, make it satisfying) — five highlights total, not fifteen. As you read, you pause twice to ask why the four laws map onto the loop rather than existing independently (elaborative interrogation), and you jot one real habit from your own life next to each law.
Immediately after the chapter (the chapter checkpoint, 5 minutes): You close the book and write three sentences from memory: what the loop is, what the four laws are, and which law you personally violate most often. You cannot fully recall the third law on the first attempt — that gap is useful information, not a failure, and you flag it for the next review.
Day 1 review: You reopen your three sentences, not the book, and try to reconstruct the four laws from the loop alone before checking. This time you get all four.
Day 7 review, in a different context: On a walk, with no notes, you try to explain the framework aloud as if to a friend who has not read the book. This is where the Feynman Technique (Strategy 6) actually happens — and where you discover you can state the four laws but cannot yet explain why "make it satisfying" matters as much as it does. That gap sends you back to the chapter for five minutes, not a re-read of the whole thing.
Day 30 review: You connect the framework to a different book — if you have also read Duhigg's The Power of Habit, you compare his cue-routine-reward loop to Clear's four laws and note where they overlap and where they don't. This cross-book connection is the highest-value review you will do, and it only works because the first three reviews already got the core framework into long-term storage.
Total added time across the whole process: roughly 20 minutes spread over a month, for a chapter most readers would otherwise finish, nod at, and forget by chapter six. The pattern generalizes to any nonfiction chapter with a nameable framework — checkpoint immediately, review at expanding intervals, and force explanation rather than recognition at each step.
Strategy 8: Let an AI Tutor Quiz You on the Source
The seven strategies above all predate the tools available in 2026. The newest addition to the retention toolkit is the conversational AI tutor — and it deserves its own entry, because it solves a specific failure mode the other strategies leave open.
The failure mode is this: when you test yourself, you write the questions, and you tend to write the questions you can already answer. Self-quizzing quietly drifts toward your comfort zone. An AI tutor that has the actual text of the chapter does not have that bias. It can ask the question you were hoping to avoid, follow up when your answer is vague, and surface the connection between chapter three and chapter eleven that you never noticed.
This is elaborative interrogation (Strategy 3) and the Feynman Technique (Strategy 6) executed by a partner instead of done alone — which matters, because doing them alone requires the discipline that fails most readers at 11pm. To use an AI tutor well:
- After each chapter, ask it to quiz you cold before you re-read anything. Recall first, verify second.
- When your answer is fuzzy, ask it to point to the exact passage you missed, then re-read only that.
- Ask it to connect the chapter to a different book you have read — cross-book synthesis is where shallow understanding gets exposed.
The key requirement is that the tutor is tied to the source. A generic chatbot answering from its training data is not testing your memory of this book; it is answering from its own. The retention gain comes specifically from being quizzed on the highlights and chapters you read. (If you are using a general assistant rather than a reading-native tool, our guide to using ChatGPT for books covers the prompts that work and the limits to watch for.)
Choosing a Retention Tool in 2026
The mistake most readers make when picking a tool is choosing on features rather than on friction. Every retention system on the market works in theory. The one that works in practice is the one whose card-making and review steps are low-friction enough that you do not abandon them in week three.
A few honest distinctions for 2026:
- General-purpose flashcard apps (Anki and similar) have the best scheduling algorithms but make you build every card by hand — which is the most common reason the habit dies. They suit defined material like vocabulary and exam facts more than books.
- Highlight resurfacers show old highlights passively. Better than nothing, but passive resurfacing is recognition, not recall — and recognition is the trap this whole guide is built to escape.
- Reading-native retention apps generate review cards from the highlights you already make, keep each card linked to its source, and pair review with active testing — see the best reading retention apps for a criteria-based comparison.
If you want a criteria-driven breakdown of every major option — Anki, RemNote, SuperMemo, Quizlet, Brainscape, Mochi, Memrise, and Chapterly — see the best spaced repetition apps in 2026. SuperMemo's own reading method, incremental reading, is worth understanding on its own before you adopt a tool built around it. For the underlying method before you pick anything, spaced repetition for readers is the place to start.
Building Your Reading System
The most important thing is consistency. A simple system you actually use beats a perfect system you abandon.
Start with these fundamentals:
- Capture highlights as you read (digital or physical)
- Process each chapter with a brief checkpoint
- Review highlights using spaced repetition
- Connect ideas across books and to your own experience
Tools That Help
While you can implement these techniques with paper and pen, technology can remove friction. Here's how the most popular options compare:
| Tool | Best For | Spaced Repetition | Active Recall | Highlights | AI Discussion |
|---|---|---|---|---|---|
| Chapterly | Readers who want full retention | Yes (built-in) | Yes (AI-powered) | Yes | Yes |
| Readwise | Highlight collectors | Yes (resurfacing) | Limited | Yes (multi-source import) | No |
| Anki | Flashcard-based memorization | Yes (FSRS algorithm) | Yes (manual cards) | No | No |
| Obsidian | PKM enthusiasts | Plugin-based | Manual | Manual | Plugin-based |
| Notion | General note-takers | No | Manual | Manual | Limited |
For detailed comparisons, see our guides to the best reading apps, Chapterly vs Readwise, the best Readwise alternatives in 2026, and the best book highlights apps. The best tool is the one you'll actually use consistently.
A Sample Reading Retention Workflow
Here's a practical daily workflow that combines the strategies above. It adds about 15-20 minutes to your reading time but can increase retention from 10% to 70%+:
During Reading (add 5 min per chapter):
- Highlight selectively—5-10 passages per chapter maximum
- Annotate highlights with brief "why this matters" notes
- Ask "why?" and "how?" as you read (elaborative interrogation)
After Each Chapter (5 minutes):
- Close the book and write a 2-3 sentence summary from memory
- Identify the single most important idea
- Note one connection to something else you know
- Write one question the chapter raised
After Finishing the Book (10 minutes):
- Use the Feynman Technique to explain the 3 most important ideas
- Export highlights to your review system
- Identify 1-3 action items to implement within the week
Daily Review (5-10 minutes):
- Review highlights from past reading via spaced repetition
- For each highlight, practice active recall before re-reading
- Note any new connections between different books
This workflow integrates spaced repetition, active recall, elaborative interrogation, and the Feynman Technique into a sustainable routine. For help building this system, see our guide to building a reading system.
Frequently Asked Questions
How long does it take to start remembering more from books?
Most readers notice a significant improvement within 2-3 books when applying these techniques consistently. The chapter checkpoint habit (Strategy 4) produces immediate results—you'll notice better recall even the same day. Spaced repetition and active recall show their full power after 2-4 weeks of consistent practice, when you start recalling details from books you read weeks ago.
Is re-reading a book ever worthwhile?
Re-reading can be valuable for truly foundational books that you want to internalize deeply, but it's one of the least efficient retention strategies for most reading. You're better off reading once with active engagement (highlighting, summarizing, questioning) plus spaced review of highlights than passively re-reading the entire book—a pattern researchers call the generation effect: making your own notes beats re-reading. If you do re-read, focus on the chapters or sections most relevant to your current thinking rather than the whole book.
Do these techniques work for audiobooks?
Yes, with adaptation. You can't annotate while listening, but you can: pause after each chapter and mentally summarize the key points, use an audiobook app with built-in note-taking to capture quick takeaways, and discuss what you heard with a friend or AI assistant. The active recall component is especially powerful for audiobooks—pausing to recall forces processing that passive listening doesn't provide.
How many highlights should I save per chapter?
Aim for 5-10 highlights per chapter. Research suggests that selective highlighting outperforms extensive highlighting because it forces you to evaluate what's truly important. If you're highlighting more than 15% of the text, you're not being selective enough. Each highlight should represent something you'd want to revisit in six months. (Side note for readers who use yellow-pink-blue-green systems: the research on whether color-coding your highlights actually helps you remember more is more cautious than the productivity-blog consensus — a single-color, selective system tends to outperform an elaborate color taxonomy you cannot maintain.) The underlying mechanism is the Von Restorff (isolation) effect: distinctive items get a memory bonus, but the bonus disappears when most of the page is highlighted — saturation eliminates the contrast the effect depends on. Whatever highlights survive your selectivity discipline, run them through successive relearning — bring each one back to a recall criterion across several spaced sessions — and the retention payoff at three months is dramatically larger than highlighting alone produces.
What's the best time of day to read for retention?
Research varies, but most studies suggest that reading during your peak alertness hours produces better comprehension and encoding. For most people, this is morning or mid-morning. However, reading before sleep also has retention benefits because sleep is when memory consolidation occurs. The worst time is when you're fatigued and likely to read passively.
Can you remember what you read without taking notes?
Yes, but it requires more deliberate mental engagement. The chapter checkpoint (Strategy 4) and the Feynman Technique (Strategy 6) work entirely in your head. Even without written notes, the act of pausing to mentally summarize, question, and connect ideas significantly improves retention compared to continuous reading. That said, written notes create external memory aids that extend your natural recall even further.
Does speed reading help or hurt retention?
Speed reading typically hurts retention. Most speed reading techniques work by reducing sub-vocalization (inner speech) and minimizing regressions (re-reading), both of which are actually helpful for comprehension and memory encoding. Research by Rayner et al. (2016) found that speed-reading programs generally sacrifice comprehension for speed. For most readers, reading at a comfortable pace with active engagement produces far better results than racing through text. For more on balancing speed and retention, see our guide on how to read more books.
Conclusion
Remembering what you read isn't about having a better memory—it's about using your memory correctly. By applying spaced repetition, active recall, and elaborative interrogation, you can retain far more from every book you read.
The investment is small: a few minutes of processing after each chapter, plus brief daily review sessions. The return is enormous: actually keeping the knowledge you work hard to acquire.
Start with one book. Apply these techniques. Watch how much more you remember. Then never go back to passive reading again.
Related Reading
- Spaced Repetition for Readers: Beyond Flashcards — A practical guide to using spaced repetition without creating thousands of flashcards.
- What Is Active Recall? The Complete Guide — Deep dive into the most effective learning technique.
- The Forgetting Curve: Why You Forget Books — The science of forgetting and how to beat it.
- The Feynman Technique for Reading — How to use teach-back to deepen understanding.
- How to Take Better Book Notes — Note-taking methods that aid memory and retention.
- How to Remember Vocabulary From Books — Specific strategies for retaining new words and concepts encountered in your reading.
- What to Do After Finishing a Book — A step-by-step process for locking in a book's ideas once you turn the last page, so the retention work you did while reading actually sticks.
- The 9 Best AI Study Tools for Readers in 2026 — The seven strategies above, graded across the AI tools that claim to deliver them.
- Self-Regulated Learning for Readers — Zimmerman's 3-phase cycle of forethought, performance, and self-reflection applied to books.
- Information Overload and Reading — How to read more without retaining less through a deliberate triage system.
- How to Improve Your Vocabulary Through Reading — A strategic guide to building vocabulary faster by reading above your level and across diverse domains.
- The Matthew Effect in Reading: Why Readers Keep Getting Better — How reading ability compounds over time and why consistent readers accelerate while inconsistent ones fall behind.
- Working Memory and Reading — Why your brain drops ideas mid-chapter and how chunking, offloading, and focus protect comprehension.
- 10 Reading Retention Tips Backed by Science — Quick, actionable techniques to remember more.
- The SQ3R Reading Method — The five-step study method that applies all these techniques in one reading session.
- The PQ4R Method — The six-step upgrade to SQ3R for dense conceptual material; adds explicit Preview and Reflect phases.
- How to Read a Difficult Book — A practical protocol for genuinely hard texts: pacing, decoding terminology, and avoiding the working-memory blowout that ends most attempts.
- How to Read Plays — Why drama is its own reading mode, and how to keep characters, scene structure, and stage directions tracked as you go.
- Reading Speed vs Comprehension Tradeoff — What 60 years of eye-tracking research actually shows about how fast you can read with full comprehension.
- How to Keep a Commonplace Book — Build a personal themed reference from your reading over years.
- Learning How to Learn: A Complete Guide for Self-Directed Readers — The meta-skills and cognitive science principles that make every other learning strategy more effective.
- 7 Active Reading Strategies That Actually Work — Evidence-based techniques that help you comprehend more and remember longer.
- How to Build a Reading System — Create a sustainable reading practice that sticks.
- Non-Fiction Reading Strategies — Specialized techniques for getting more from non-fiction.
- The Serial Position Effect: Why You Remember Some Parts of Books More Than Others — Why readers naturally remember chapter openings and endings but lose the middle, and what to do about it.
- How to Memorize a Book: A 7-Step System for Locking In Key Ideas — The complete system for converting books from vague impressions into durable, accessible knowledge.
- How AI Reading Assistants Are Changing Book Learning — How to use AI to deepen your reading comprehension after every chapter.
- The Pretesting Effect: Why Guessing Before You Read Makes You Learn More — The counterintuitive science showing that wrong guesses before reading prime your brain to encode the correct answers more deeply.
- How to Analyze a Book: A Practical Framework for Going Beyond Summary — Five-step framework to identify arguments, evaluate evidence, surface assumptions, and form your own reasoned position.
A Deeper Cut: Why Most Retention Advice Fails Even When Followed
One pattern worth naming directly: readers who know these techniques still forget books. The failure mode is almost never a knowledge gap about the techniques themselves. It is that the fluency illusion — the mistaken sense that smooth reading equals learning — quietly cancels them out. A reader finishes a chapter, feels they "got it," and skips the retrieval step because it seems unnecessary. Robert Bjork's research on retrieval strength versus storage strength shows precisely why this feeling is unreliable: what feels learned and what is learned are nearly uncorrelated in the moment. The retention techniques above work, but only if you apply them when your gut says you do not need them. That is the uncomfortable discipline that separates readers who retain from readers who merely underline.
A 2023 replication of the Karpicke and Roediger retrieval-practice studies extended this finding to ebook reading specifically. Students who used highlighting as their primary study tool scored at the same baseline as students who did not study at all — both roughly 40% on a one-week delayed test. Students who used any form of active retrieval, even brief self-quizzing, scored between 65% and 78%. The gap between "feels like studying" and "is studying" remains the largest variable in reading retention research.
Ready to implement these strategies automatically? Chapterly combines AI-powered chapter discussions with spaced repetition review, making it easy to remember everything you read. Try it free.