Start free trial
Trending New Genres Authors
EnglishEnglish
EspañolSpanish
简体中文Chinese
繁體中文Chinese (Traditional)
FrançaisFrench
DeutschGerman
日本語Japanese
PortuguêsPortuguese
ItalianoItalian
한국어Korean
РусскийRussian
NederlandsDutch
العربيةArabic
PolskiPolish
हिन्दीHindi
Tiếng ViệtVietnamese
SvenskaSwedish
ΕλληνικάGreek
TürkçeTurkish
ไทยThai
ČeštinaCzech
RomânăRomanian
MagyarHungarian
УкраїнськаUkrainian
IndonesiaIndonesian
DanskDanish
SuomiFinnish
БългарскиBulgarian
עבריתHebrew
NorskNorwegian
HrvatskiCroatian
CatalàCatalan
SlovenčinaSlovak
LietuviųLithuanian
SlovenščinaSlovenian
СрпскиSerbian
EestiEstonian
LatviešuLatvian
فارسیPersian
മലയാളംMalayalam
தமிழ்Tamil
اردوUrdu
Searching...
SoBrief
Deep Medicine

Deep Medicine

How Artificial Intelligence Can Make Healthcare Human Again
The best use of medical AI may be giving doctors time to understand each patient.
by Eric J. Topol 2019 341 pages
4.01
2k+ ratings
Listen
Immersive
V2.3
Amazon Kindle Audible
Try Full Access for 3 Days
Unlock listening & more!
Continue

Key Takeaways

Make AI the assistant that gives clinicians time to care

The human purpose. Topol calls for medicine that combines a fuller picture of each person, machine analysis, and a stronger bond between patient and clinician. His own knee replacement shows what is at stake: a standard, forceful rehabilitation protocol worsened a rare scarring complication, while a therapist who understood his history and stayed in contact helped him recover. Better data might have flagged the risk, but attentive care made the difference.

Machines can contribute. In a newborn with relentless seizures, rapid genome analysis and software that extracted clinical features helped identify a rare metabolic disorder. A change in diet stopped the seizures. The promise is not to replace caregivers, but to use machines for data-heavy work so people can make better decisions together.

Analysis

The argument treats technology as a means to a human end, not a measure of progress in itself. That distinction resembles a core question in organizational design: when automation frees capacity, who receives the benefit? In healthcare, efficiency can become more appointments and tighter schedules, or more attention for each patient. The technical improvement alone does not decide. That requires incentives and professional norms that value time spent understanding a person, not just throughput.

Treat the patient, not the first plausible pattern

A rushed diagnosis. A patient named Robert was told that a small opening between heart chambers caused his brief stroke and that it should be plugged. Topol questioned that conclusion and arranged longer heart-rhythm monitoring. It revealed silent episodes of atrial fibrillation, a more plausible stroke cause, and changed the plan from an implant to blood-thinning medication.

A system-wide problem. The book cites an estimate of roughly 12 million significant diagnostic errors in the United States each year. Brief visits, missing records, incomplete testing, and snap judgments can compound one another. Robert’s case was not simply a mistake by one clinician. It showed how shallow contact can yield a shallow account of what is happening, followed by an unnecessary procedure.

Analysis

Robert’s case illustrates how diagnosis is a process of revising competing explanations, not matching one symptom to one disease. In decision science, a striking clue can dominate attention even when it is common in people without the illness. More data help only if the question is framed well and alternatives remain open. The case also resists a simple technology fix: longer monitoring revealed the cause, but a clinician still had to decide which finding mattered.

Let algorithms read patterns, but keep humans responsible for context

A narrow strength. Deep learning is a form of machine learning that uses multiple layers of computation to find patterns in examples. It can be especially useful for images and signals. Studies described in the book found that algorithms could match or exceed specialists on specific tasks such as classifying skin lesions or spotting abnormalities in scans. These are bounded comparisons, not proof that a machine can practice medicine independently.

Context changes meaning. A scan-reader may excel at a defined classification while missing why the scan was ordered, what else is visible, or how the result fits a person’s history. Topol argues that radiologists and pathologists can shift from repetitive interpretation toward checking recommendations, explaining results, and discussing care directly with patients.

Analysis

The useful comparison is not machine versus doctor, but task versus whole job. A single test can be benchmarked against labeled examples; clinical care requires choosing the test, interpreting uncertainty, and discussing consequences. This distinction helps explain why a strong score on a narrow dataset can coexist with weak performance in practice. Human oversight also needs substance. A hurried clinician who merely approves an output is not a meaningful safeguard.

Feed algorithms representative data, then test them in real clinics

Bad inputs mislead. AliveCor and Mayo Clinic first trained an algorithm to estimate blood potassium from electrocardiograms. The initial results were poor. The dataset had excluded hospitalized patients, who were more likely to have dangerous potassium levels, and the team had focused on one part of the tracing. With a broader set of patients and the full signal, performance improved; a later prospective test in dialysis patients supported further evaluation.

A score is not deployment. Topol distinguishes retrospective analysis of stored records from prospective testing with patients in ordinary care. Data labels must be accurate, the sample must reflect intended use, and performance can change as populations and practice change. A promising result on yesterday’s records is evidence to investigate, not proof that a tool is safe or useful everywhere.

Analysis

This is a lesson in measurement as much as computation. The algorithm did not simply need more examples; it needed examples that represented the cases it was supposed to handle, paired with measurements taken at the right time. The distinction resembles the difference between a laboratory instrument and a field instrument: controlled conditions can hide failure modes. Prospective testing exposes workflow, patient, and data complications that retrospective benchmarks cannot.

Ask whether a test helps this person, not merely finds disease

Detection can harm. Screening finds abnormalities, but some would never cause illness. In the book’s example, screening 10,000 women in their fifties annually for a decade was estimated to prevent five breast-cancer deaths, while more than 6,000 women would receive at least one false-positive result. In South Korea, widespread thyroid screening drove diagnoses sharply upward without a corresponding fall in deaths from the disease.

Start with individual risk. A result means different things depending on how likely the person was to have the condition before testing. Topol argues against ordering tests indiscriminately, because false alarms can trigger anxiety, invasive procedures, and treatment without meaningful benefit. A useful test is not simply sensitive or technologically impressive. It should help a particular patient make a better decision.

Analysis

This insight joins two questions often separated in medicine: whether a test detects something accurately, and whether detecting it improves the patient’s life. Screening can appear successful when diagnosis rates rise, even if serious outcomes do not improve. That makes outcomes, not activity, the essential measure. The examples do not mean screening is always wrong. They make a stronger case for matching the test to a person’s risk and being candid about false positives and overdiagnosis.

Data carry history. Algorithms learn from the records and examples they receive. If those examples underrepresent certain groups, or reflect past discrimination, the resulting tool can repeat or intensify the imbalance. Topol highlights the particular risk of medical algorithms trained on populations that do not represent the people later assessed, including skin-cancer systems developed with few patients of color.

Useful tools still need safeguards. A UK health service shared records from 1.6 million people with DeepMind without explicit consent to develop an app for detecting kidney injury. The app helped staff find relevant information faster, but regulators concluded the data transfer was not lawful. The case shows why patient benefit does not erase the need for transparency, permission, security, and clear limits on data use.

Analysis

Fairness is not a cosmetic adjustment after a model is built. It concerns whose experiences count as evidence and who bears the cost when an error occurs. Consent matters for the same reason: health data describe intimate aspects of a life, not interchangeable raw material. At the same time, privacy rules should not make beneficial research impossible. The hard task is building accountable arrangements that permit useful analysis without treating access to data as automatic.

Use mental-health bots to widen access, not impersonate therapists

Some people disclose more. In a study of 239 participants, people interviewed by a virtual character shared more when they believed it was computer-controlled than when they thought a human was operating it. The book connects this result to mental-health tools such as chatbots that offer structured support and cognitive behavioral therapy exercises. A small trial of Woebot, a conversational program, found promising signs of engagement and reduced symptoms among college students.

Reach is not a cure. Depression affects hundreds of millions of people, and many receive no treatment. Digital tools could make support easier to reach, especially where trained professionals are scarce. But the studies described are limited, and sensitive conversations create serious privacy concerns. A bot may offer a useful first step or supplement, but it cannot be assumed to provide the trust, judgment, or care of a human professional.

Analysis

The disclosure finding is revealing because it frames privacy not only as protection from outsiders, but as freedom from anticipated judgment. A machine can lower that social barrier, potentially helping people begin a difficult conversation. Yet apparent comfort can be misleading if users do not know how their words are stored or used. Access and confidentiality must be designed together. The strongest role for a bot is a defined, safe support function with a route to human help.

Treat personalized diets as promising estimates, not proven prescriptions

People respond differently. Researchers at Israel’s Weizmann Institute monitored blood glucose in 800 people without diabetes, collecting more than 1.5 million measurements alongside information about food, activity, sleep, and gut microbes. A machine-learning model used 137 factors to predict an individual’s glucose response. In a small follow-up comparison, personalized meal plans improved post-meal glucose measurements.

A surrogate is not health itself. In a bread study, people differed in their responses to white and sourdough bread, even when the average response across the group looked similar. The research suggests that standard dietary advice can hide individual variation. But improved glucose readings do not by themselves show that a diet prevents diabetes or extends life. The book treats personalized nutrition as an intriguing early result, not a settled consumer guarantee.

Analysis

The work challenges the assumption that a population average is a prescription for every person. It also demonstrates how combining continuous measurements with behavioral and biological data can reveal patterns that a food list alone would miss. Still, an accurate prediction is not automatically a useful intervention. A recommendation needs to be practical, safe alongside a person’s other conditions, and shown to improve outcomes that matter. Personalization raises the standard of evidence rather than lowering it.

Keep your health data under your control, not locked in systems

The record is incomplete. Topol argues that current electronic health records were built largely around billing and administration, not as a complete, accurate history for patients. Records may be scattered across providers, copied forward with errors, or disconnected from sensor readings, genetic information, and daily life. A virtual medical assistant could only offer useful guidance if it could draw on relevant information over time.

Control is part of safety. The book calls for people to have access to and authority over their health information, with strong protections against theft, hidden sharing, or use by employers and insurers to penalize them. Estonia is offered as an example of a system where individuals control access to their records. Collecting more data without trustworthy governance would deepen risk, not necessarily improve care.

Analysis

The deeper issue is not simply data ownership as a legal label. It is whether a patient can understand who has access, correct errors, and refuse uses that conflict with personal interests. A comprehensive record can support continuity, but it can also make surveillance more powerful. Trustworthy data systems therefore need both interoperability and boundaries. The book’s vision is most persuasive when patient control is treated as a condition for participation, not as a technical feature added after collection.

Spend AI's efficiency dividend on listening, touch, and longer visits

Time changes care. Topol argues that automation should reduce clerical work and repetitive analysis, giving clinicians more time with patients. One study he cites found that longer home-health visits were associated with fewer hospital readmissions. Yet efficiency can be redirected toward seeing more patients rather than caring more deeply for each one. The technology cannot choose how the saved time is spent.

Presence is clinical work. Patients need clinicians who listen to their account, examine them, explain results, and acknowledge fear or grief. The book notes that doctors interrupt patients, on average, just eighteen seconds after they begin speaking. It also argues that empathy can be cultivated through education and practice. AI may make knowledge more available, but trust and the experience of being cared for remain human responsibilities.

Analysis

This is the book’s institutional challenge: productivity gains do not automatically become better care. They follow the incentives of the organizations that control staffing, scheduling, and reimbursement. The proposed human dividend therefore depends on governance, not goodwill alone. The emphasis on listening and examination also guards against a subtle error: treating empathy as a pleasant extra after the technical work. In medicine, communication is part of how evidence is gathered and decisions become usable.

Analysis

Deep Medicine is a science and medicine argument told through clinical stories, research examples, and warnings about implementation. Its central move is to redefine AI’s highest promise. The prize is not a machine that replaces a doctor, but a division of labor in which computation handles scale and pattern recognition while clinicians recover time for judgment, explanation, and relationship. The book’s strongest evidence comes from cases where data changes a decision, such as rapid genomic analysis of a critically ill newborn or prolonged heart monitoring that uncovers silent atrial fibrillation. Its most useful caution is that narrow technical accuracy does not establish clinical value. Training data can be distorted, retrospective tests can flatter performance, and an early diagnosis can create harm when it does not change outcomes.

The argument also identifies a political economy problem. Medical systems may capture productivity gains as higher throughput, while patients and clinicians continue to experience hurried encounters. Thus, the author’s proposed remedy depends on institutions changing how they reward care, not just on building better software. The call for patient control of data follows the same logic: information is useful only within a relationship of legitimate access and accountability.

Some evidence remains preliminary within the book’s own account. Personalized diets show predictive promise for glucose response, not established prevention of disease. Mental-health applications may improve access, but small studies do not settle effectiveness, privacy, or the quality of support. These limits strengthen rather than weaken the broader thesis. Deep medicine is best read as a conditional program: advanced analytics can help, provided that systems validate tools in real practice, monitor unequal effects, protect patients’ information, and deliberately return saved time to care. Its enduring question is not whether machines can perform medical tasks, but what kind of healthcare institutions will use their capabilities to create.

Summary in 30 Seconds
AI can identify patterns in scans, signals, and records, but its findings need clinical context and real-world testing with representative patients. An estimated 12 million significant diagnostic errors occur annually in the United States. Screening can also harm: one estimate found that screening 10,000 women in their fifties yearly for a decade would prevent five breast-cancer deaths while more than 6,000 received at least one false positive. Health data require consent and safeguards. Automation helps patients most when saved time goes to listening and explanation.
Contains spoilers

Last updated:

Report Issue

Review Summary

4.01 out of 5
Average of 2k+ ratings from Goodreads and Amazon.

Deep Medicine explores AI's potential to transform healthcare by improving diagnosis, freeing up doctors' time, and restoring the human touch in medicine. Reviewers praised Topol's comprehensive overview of AI applications in healthcare and his balanced perspective on its promises and limitations. Many appreciated his vision of AI enhancing rather than replacing human doctors. Some readers found the writing style dry or technical at times. Overall, the book was seen as an important and timely examination of AI's role in shaping the future of medicine.

Your rating:
4.47
685 ratings
Want to read the full book?

Glossary

Deep medicine

Data-rich, human-centered care

Topol’s vision of healthcare that combines a more complete understanding of each person, machine learning to analyze complex information, and stronger human relationships between patients and clinicians. Its goal is not simply faster diagnosis or lower cost. Technology should support more individualized decisions while helping restore attention, empathy, and trust.

Shallow medicine

Fragmented, rushed care

Topol’s term for care shaped by brief encounters, incomplete information, weak patient-clinician connection, and decisions made without enough context. It can produce missed diagnoses, unnecessary tests, and treatments that offer little benefit. The term describes a system-level pattern, not merely an individual clinician’s mistake.

Deep phenotyping

Broad, longitudinal health data

A detailed characterization of an individual using many kinds of information, potentially including medical and family history, social and behavioral context, anatomy, physiology, genetics, and environment. Topol emphasizes that the data should be both broad and collected over time, because health measures change and isolated snapshots can miss important patterns.

Digital phenotyping

Behavior measured through devices

The collection and analysis of digital traces that may reflect a person’s behavior or mental state. Examples in the book include speech patterns, typing rhythm, smartphone activity, facial expression, and sensor measurements such as heart rate or sleep-related movement. These signals may aid assessment, but their accuracy and clinical meaning require validation.

Augmented individualized medical support

Personalized clinical decision support

Topol’s proposed alternative to conventional clinical decision support systems. It would use a patient’s broad, individual information, together with relevant medical evidence, to help clinicians make decisions tailored to that person rather than relying only on population averages. The proposal depends on comprehensive, reliable records and tools that fit clinical work.

Digital twins

Closely matched patient comparisons

In Topol’s usage, digital twins are other people whose measured characteristics resemble a patient across relevant demographic, biological, physiological, and anatomical dimensions. Their treatments and outcomes could inform decisions for the patient being assessed. The idea depends on large, representative datasets and careful safeguards against privacy loss and misleading matches.

Glucotypes

Patterns of glucose response

Categories used in research described by Topol to group recurring patterns in how people’s blood glucose changes after eating. The work suggests that glucose responses can vary substantially among individuals, even after similar meals. Glucotypes are a way to characterize this variation, not a complete measure of health or proof that a particular diet prevents disease.

Therapeutic illusion

Overestimating treatment benefit

A term discussed by Topol for clinicians’ tendency to believe that the tests or procedures they provide are more beneficial than evidence supports. A clinician may interpret a treatment as successful because it was performed or because the patient improved, even when outcomes do not demonstrate its value. This can help sustain unnecessary care.

FAQ

What's Deep Medicine about?

  • Focus on AI in Healthcare: Deep Medicine by Eric J. Topol explores how artificial intelligence (AI) can enhance the human aspects of healthcare, making it more personalized and empathetic.
  • Integration of Data: The book discusses integrating vast amounts of medical data to create a comprehensive view of individual health, aiming to tailor treatments to each patient's unique needs.
  • Future of Medicine: Topol envisions a future where AI and deep learning tools work alongside healthcare professionals to improve patient outcomes and enhance the overall experience of care.

Why should I read Deep Medicine?

  • Insightful Perspective: The book provides a unique viewpoint on the intersection of technology and healthcare, emphasizing the need for a balance between AI capabilities and human empathy.
  • Practical Applications: Topol illustrates how AI can be applied in real-world scenarios, such as diagnosing diseases and personalizing treatment plans, making it relevant for healthcare professionals and patients.
  • Addressing Current Issues: The book tackles pressing issues in healthcare, such as misdiagnosis and burnout among healthcare providers, making it a timely read.

What are the key takeaways of Deep Medicine?

  • AI as a Tool, Not a Replacement: Topol emphasizes that AI should enhance, not replace, the human elements of healthcare, fostering deeper connections between patients and providers.
  • Importance of Deep Learning: The book highlights the role of deep learning in analyzing complex medical data, leading to more accurate diagnoses and better treatment outcomes.
  • Need for Empathy in Medicine: A recurring theme is the necessity of empathy in patient care, with the human touch remaining irreplaceable in understanding and addressing patients' needs.

What are the best quotes from Deep Medicine and what do they mean?

  • “The secret of the care of the patient is caring for the patient.”: This quote underscores the importance of empathy in healthcare, suggesting that good medical practice lies in truly understanding and caring for patients.
  • “We want the physician—a caring individual and not a machine—to give us time.”: Topol emphasizes that patients desire a personal connection with their healthcare providers, highlighting the need for time and attention in medical interactions.
  • “AI in medicine isn’t just a futuristic premise.”: This quote reflects the current reality of AI applications in healthcare, indicating that these technologies are already being utilized to improve patient care and outcomes.

What is the concept of "deep medicine" as defined in Deep Medicine?

  • Definition of Deep Medicine: Topol defines "deep medicine" as a model that integrates advanced AI tools with a deep understanding of individual patients, enhancing human aspects of care.
  • Three Components: The model consists of deep defining of individuals, deep learning for accurate diagnostics, and deep empathy to foster meaningful patient-provider relationships.
  • Transforming Healthcare: The concept advocates for a transformation in healthcare where technology and human connection coexist, leading to a more personalized and effective healthcare system.

How does AI improve medical diagnosis according to Deep Medicine?

  • Pattern Recognition: AI excels at recognizing patterns in large datasets, leading to more accurate diagnoses, such as identifying abnormalities in medical images.
  • Data Integration: AI integrates various types of data to provide a comprehensive view of a patient's health, enhancing diagnostic accuracy and treatment planning.
  • Real-World Applications: Topol provides examples of AI applications in diagnosing conditions like skin cancer and heart disease, demonstrating improved patient outcomes.

What are the limitations of AI in healthcare as discussed in Deep Medicine?

  • Black Box Problem: A significant limitation is the "black box" nature of many AI algorithms, where the decision-making process is not transparent, hindering trust.
  • Bias and Inequities: AI systems can perpetuate existing biases present in the training data, leading to inequitable healthcare outcomes.
  • Dependence on Data Quality: The effectiveness of AI is heavily reliant on the quality and comprehensiveness of the data used for training, affecting predictions and diagnoses.

How does Deep Medicine address the issue of empathy in healthcare?

  • Empathy as Essential: Topol argues that empathy is a critical component of effective healthcare, emphasizing that technology should enhance, not replace, the human connection.
  • Impact of Burnout: The book discusses how technology can contribute to physician burnout, affecting patient care quality, and stresses restoring empathy in medicine.
  • Training for Emotional Intelligence: Topol suggests prioritizing emotional intelligence and communication skills in medical training to cultivate healthcare providers who can effectively connect with patients.

What role does data play in the future of healthcare as envisioned in Deep Medicine?

  • Big Data Utilization: The book highlights the importance of utilizing big data to inform medical decisions and improve patient outcomes.
  • Personalized Medicine: By integrating diverse data sources, healthcare can become more personalized, allowing for tailored treatment plans.
  • Continuous Learning: Topol envisions AI systems continuously learning from new data, leading to ongoing advancements in medical knowledge and practice.

How does Deep Medicine propose to change the healthcare system?

  • Integration of AI and Human Care: Topol advocates for a healthcare system that seamlessly integrates AI tools with compassionate human care.
  • Empowering Patients: The book emphasizes empowering patients by providing access to their health data and involving them in decision-making processes.
  • Reforming Medical Education: Topol calls for reforms in medical education to prioritize both technical skills and emotional intelligence.

How does Deep Medicine address the role of AI in diagnostics?

  • Enhanced Diagnostic Accuracy: AI can analyze vast amounts of data to improve diagnostic accuracy, particularly in fields like radiology and pathology.
  • Predictive Analytics: The book highlights AI's use in predicting patient outcomes, allowing for more proactive and personalized care.
  • Integration with Clinical Practice: Topol emphasizes the need for AI tools to complement rather than complicate healthcare providers' work.

What ethical considerations does Deep Medicine raise regarding AI in healthcare?

  • Data Privacy and Security: Topol stresses the importance of protecting patient data as AI systems become more prevalent in healthcare.
  • Bias in Algorithms: The book discusses potential biases in AI algorithms, calling for transparency and accountability to ensure equitable treatment.
  • Human Oversight: Topol argues for human oversight to ensure ethical and compassionate care, warning against over-reliance on technology.

About the Author

Eric J. Topol is a renowned cardiologist, geneticist, and digital medicine researcher. He is the founder and director of the Scripps Research Translational Institute and a professor of molecular medicine at Scripps Research. Topol is known for his work in personalized medicine and the use of genomic and digital technologies in healthcare. He has authored several influential books on the future of medicine and technology, including "The Patient Will See You Now" and "The Creative Destruction of Medicine." Topol is also an active voice on social media, particularly Twitter, where he shares insights on healthcare, technology, and current medical issues. His expertise in both medicine and technology makes him a leading figure in discussions about the intersection of AI and healthcare.

Download PDF

To save this Deep Medicine summary for later, download the free PDF. You can print it out, or read offline at your convenience.
Download PDF
File size: 0.26 MB     Pages: 20

Download EPUB

To read this Deep Medicine summary on your e-reader device or app, download the free EPUB. The .epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.
Download EPUB
File size: 1.38 MB     Pages: 26
Want to read the full book?
Follow
Listen
Home
Swipe
Library
Get App
Try Full Access for 3 Days
Listen, bookmark, and more
Compare Features Free Pro
📖 Read Summaries
Read unlimited summaries. Free users get 3 per month
🎧 Listen to Summaries
Listen to unlimited summaries in 40 languages
—
❤️ Unlimited Bookmarks
Free users are limited to 4
—
📜 Unlimited History
Free users are limited to 4
—
📥 Unlimited Downloads
Free users are limited to 1
—
Risk-Free Timeline
Today: Get Instant Access
Listen to full summaries of 26,000+ books. That's 12,000+ hours of audio!
Day 2: Trial Reminder
We'll send you a notification that your trial is ending soon.
Day 3: Your subscription begins
You'll be charged on Nov 24,
cancel anytime before.
Consume 2.8× More Books
2.8× more books Listening Reading
Our users love us
1M+ readers
Trustpilot Rating
TrustPilot
4.6 Excellent
This site is a total game-changer. I've been flying through book summaries like never before. Highly, highly recommend.
— Dave G
Worth my money and time, and really well made. I've never seen this quality of summaries on other websites. Very helpful!
— Em
Highly recommended!! Fantastic service. Perfect for those that want a little more than a teaser but not all the intricate details of a full audio book.
— Greg M
Save 62%
Yearly
$119.88 $44.99/year/yr
$3.75/mo
Monthly
$9.99/mo
Start a 3-Day Free Trial
3 days free, then $44.99/year. Cancel anytime.
Unlock a world of fiction & nonfiction books
26,000+ books for the price of 2 books
Read any book in 10 minutes
Discover new books like Tinder
Request any book if it's not summarized
Read more books than anyone you know
#1 app for book lovers
Lifelike & immersive summaries
30-day money-back guarantee
Download summaries in EPUBs or PDFs
Cancel anytime in a few clicks

Why one summary?

Scanner
Find a barcode to scan

⭐
⭐
⭐
We have a special gift for you
Open
38% off
DISCOUNT FOR YOU
⭐
⭐
⭐
$79.99
$49.99/year
Only $4.16 per month
Continue
2 taps to start, super easy to cancel
Settings
General
Widget
Loading...