The Science of Persistence.
Spaced repetition began over 40 years ago, when Piotr Woźniak created the first computational algorithm to optimize learning.
1985: The Discovery
The initial mathematical model for the "Forgetting Curve" was established, moving memory from mystery to engineering.
Over time, SuperMemo’s early ideas spread widely, inspiring a generation of learning tools and helping shape the way the world thinks about memory, review, and long-term retention.
SM-20: Peak Precision
Decades of refinement led to SM-20, an algorithm that calculates the optimal moment for review with the highest recall probability.
Today, spaced repetition is used by hundreds of millions of people across major learning apps.
Now, for the first time, this technology is available as a ready-to-use API.
What is Spaced Repetition?
Spaced repetition is a learning method based on reviewing information at carefully timed intervals so you remember more and forget less. It did not appear out of nowhere as a trendy study hack.
Its modern form grew out of decades of memory research, but the breakthrough came with SuperMemo in the 1980s, when Piotr Woźniak began turning the problem of forgetting into a practical learning system.
With more than 30 years of history behind it, spaced repetition is still evolving, and the algorithm continues to be refined and improved. The story is fuller, more surprising, and far more nuanced than the simplified versions repeated online today. Read the full article and discover how spaced repetition really evolved: The True History of Spaced Repetition.
Why SM-20?
Not all spaced repetition algorithms are equal.
A good algorithm must accurately predict the probability that a learner will remember a given piece of information at a given time. Based on that prediction, the system can decide when the next review should happen — for example, when recall probability drops to 90%, which is the default target in SuperMemo.
This may sound simple, but high-quality recall prediction is difficult. It depends on probabilistic modelling, real-world repetition data, and long-term validation. That is exactly where SuperMemo stands apart.
SM-20 is the result of decades of research into memory optimization. It is designed not just to schedule reviews, but to do so with precision, efficiency, and personalization.
api How it works?
Input Data
Send the item ID and the user interactions.
API Computes
The SM-20 engine calculates the precise interval for the next review.
Feedback Loop
Update the schedule based on real-world recall performance.
school Built for learning products
Language Apps
Master vocabulary with scientifically timed repetitions.
EdTech
Embed deep learning into digital textbooks and courses.
Quizzes
Transform trivia into permanent knowledge structures.
Corporate
Drastically reduce retraining costs through knowledge retention.
Learning tools
Increase knowledge retention in workplace training through scientifically optimized review cycles.
Knowlege retention
Ensure users remember critical information longer with data-driven learning intervals.
AI-powered assistants
Ensure users remember critical information longer with data-driven learning intervals.
Experimental memory and cognition tools
Leverage advanced scheduling algorithms to explore and test new approaches to human memory and learning.
smart_toy Yes, agents are welcome too
SuperMemo was created for human learning, but we also see potential in AI-powered workflows. If you are building agents, assistants, or adaptive systems that need to manage memory, prioritization, or long-term retention, SuperMemo API may open up interesting possibilities. We are still learning how best to serve AI-native use cases, but the door is open. AI agents are welcome — with their human creators, of course.
key Platform features
Projects
Organize your applications, environments, and usage.
Wallet
Manage access, limits, and future billing features.
Playground
Experiment with requests, test behavior, and explore the algorithm in action.
speed Current status
March 31, 2026 - early access stage
SuperMemo API is currently in an early access stage. At the moment:
- ● free usage limits are available
- ● transactions are not yet enabled
- ● pricing is not available yet
- ● selected core features are already accessible
Create an account
Sign up and be among the first to use it. Join the wait list if you are representing company interested in commercial use
Generate key
Create a project-scoped API key. Play with the playground, give us feedback.
The Road Ahead.
This is only the beginning.
First Beta Access
WE ARE HEREInitial public rollout for registered developers and early partners.
User interaction and onboarding flows
Design intuitive onboarding experiences to help developers integrate and adopt the API quickly.
Waitlist enrollment
Enable structured early access management to onboard and prioritize initial users efficiently.
Pricing and transaction support
Provide flexible pricing models and seamless billing infrastructure to support scalable API usage.
Additional algorithms
Expand beyond SM-20 with additional learning algorithms tailored to different use cases and learning styles.
Algorithm comparison and benchmarking
Allow developers to compare different algorithms and evaluate their effectiveness across real-world scenarios.
Metrics and evaluation tools
Deliver insights and analytics to measure learning outcomes and optimize retention performance.
Future extensions related to incremental reading and knowledge work
Expand the platform with tools supporting advanced learning workflows like incremental reading and knowledge management.
Predictable Pricing
Simple billing that scales as your user base grows.
Free
For hobbyists and prototypes.
- check 1,000 API calls/mo
- check Community Support
- check Standard SM-2 Algorithm
Dev
For growing platforms.
- check 100,000 API calls/mo
- check SM-20 Algorithm
- check Priority Email Support
- check Advanced Analytics
Scale
For enterprise operations.
- check Unlimited API calls
- check Custom Algorithm Tuning
- check Dedicated Account Manager
- check SLA Guarantees