Knowledge tracing software for personalized education
Since Jul 8Updated Nov 14, 2026Based on 53 AI answers
Reviewed by Dimitry Apollonsky
Brands AI mentions
Sophisticated knowledge-space modeling for mathematics and STEM mastery.
Also mentioned rarely: Carnegie Learning, EduStudio, Knewton Alta · knewton.com, Wiley, EduData, Riiid, Area9 Lyceum, LearnLens, NeuroLMS, Realizeit, AccelIQ, Carnegie Learning MATHia, IXL Learning, Learning Pool, McGraw Hill Education, PrepSt, PyTorch, Quantiphi, Google Gemini · tensorflow.org, Obi, Wissero, Adaptemy, AWS, CodeLeap, CYPHER Learning, Machine Learning for Science, OpenAI, Open edX, Akash Network, Anthropic, Amazon SageMaker, Microsoft Azure, University of California, Berkeley, Classting, Google Cloud, Google Cloud Gemini Enterprise Agent Platform, ClueSora, Cognaa, CT Corp., Edmentum, Google, Google Gemini · google.com, Interlaza, JoraIQ, Khan Academy, Khanmigo, Learning Commons, PSI-KT (Predictive, Scalable, and Interpretable Knowledge Tracing), MakeMyTest, MemoryLab, Motiva Education, Nebular, Origin LMS, Raptor Maps, Renaissance Learning, Scikit-learn, SciQuiry, Skillwell, Socra, Sonch AI, ThinkStem AI, Thinkster, UniKT
- pyBKTChatGPTGoogle AI Mode
- pyKTChatGPTGoogle AI Mode
- ALEKSChatGPTGoogle AI Mode
- EduKTMChatGPTGoogle AI Mode
- Knewton AltaChatGPTGoogle AI Mode
Questions and answers
ChatGPTNov 9, 2026
If you're building an AI-powered personalized education platform, there are several providers offering Knowledge Tracing (KT) algorithms and technologies.
I would divide them into two categories:
- Technology providers: Offer algorithms, SDKs, or libraries that you can integrate into your own AI platform.
- Commercial EdTech providers: Offer ready-made adaptive learning engines that track student mastery and personalize learning paths.
Knowledge Tracing estimates what a student knows, how their knowledge evolves through practice, and which concepts they need to learn next.
1. Open-source Knowledge Tracing providers
These are particularly interesting if you want to build your own AI education solution.
1. pyKT
(Strong recommendation)
- Python and PyTorch-based toolkit.
- Supports DKT, SAKT, AKT, DKVMN and other models.
- Useful for training, comparing and deploying custom learner models.
- Best for building your own AI-powered education platform.
2. pyBKT
- Implements Bayesian Knowledge Tracing.
- Estimates the probability that a student has mastered a skill.
- More interpretable than many deep-learning alternatives.
- Particularly useful for an initial prototype.
3. EduStudio
- Open-source educational data-mining framework.
- Includes multiple knowledge tracing approaches.
- Useful for benchmarking algorithms and experimenting with different learner models.
2. Commercial companies offering Knowledge Tracing or adaptive learning
These companies are worth evaluating if you prefer buying an existing technology rather than developing the entire engine.
4. CENTURY Tech
(Explicit Knowledge Tracing)
- Uses a knowledge tracing model, knowledge graphs and learner activity graphs.
- Identifies strengths, gaps and misconceptions.
- Recommends the next learning activity.
- Worth investigating for K–12 and institutional education.
5. Knewton Enterprise / Wiley
- Adaptive learning infrastructure.
- Learner analytics and adaptive recommendations.
- Designed to support organizations building digital learning products.
- Particularly interesting if you want an external technology partner.
Developer documentation · Knewton Alta ·
6. ALEKS / McGraw Hill
- Uses Knowledge Space Theory.
- Maps what learners know and what they are ready to learn.
- Strong mathematics and STEM applications.
- Primarily a complete adaptive learning product rather than a standalone KT API.
7. Area9 Lyceum
- Offers adaptive learning through Rhapsode.
- Evaluates knowledge and learner confidence.
- Personalizes content and learning pathways.
- Particularly relevant for professional and corporate education.
8. Squirrel AI Learning
- Fine-grained knowledge mapping.
- Adaptive tutoring and learner modeling.
- Personalized recommendations based on student performance.
9. Carnegie Learning
- MATHia uses cognitive modeling and adaptive instruction.
- Provides detailed skill-level feedback.
- Has roots in intelligent tutoring system research.
3. Comparison: Which providers should you evaluate?
Provider Technology Best application pyKT Deep KT algorithms Custom AI platform pyBKT Bayesian KT Mastery estimation EduStudio Multiple KT models Research and testing CENTURY Tech Proprietary KT Personalized education Knewton Adaptive engine Platform integration ALEKS Knowledge Space Theory Math and STEM Area9 Adaptive learner modeling Workforce learning Squirrel AI Adaptive learner modeling Personalized tutoring Carnegie Learning Cognitive modeling Mathematics One important distinction: commercial adaptive-learning platforms do not necessarily expose their underlying knowledge tracing algorithms through a standalone API. You should verify API access, licensing, data ownership and integration rights before procurement.
4. Recommended architecture for your AI education platform
If I were designing this system, I would separate knowledge assessment from learning-path generation.
1. Student interaction data
Answers · Assessments · Attempts · Time spent
2. Knowledge Tracing Engine
pyKT / pyBKT / Commercial KT engine
3. Learner Knowledge Profile
Mastery probabilities · Knowledge gaps · Confidence
4. Learning Path Recommendation Engine
Prerequisites · Next-best concept · Difficulty · Sequencing
5. Generative AI Tutor
Personalized lessons · Exercises · Explanations · Feedback
This separation is important: Knowledge Tracing estimates what a student knows. The recommendation engine decides what the student should learn next. The LLM generates the educational experience.
5. My recommendation
I would prioritize these three for an initial technical evaluation:
- pyKT: If you want to develop your own proprietary AI education engine.
- pyBKT: If you want an interpretable baseline and relatively straightforward implementation.
- Knewton Enterprise: If you want to investigate commercial adaptive-learning infrastructure and integration.
CENTURY Tech is also particularly interesting because its published architecture explicitly combines knowledge tracing, knowledge graphs and personalized recommendations.
6. A question that will help me narrow the vendor shortlist
Are you planning to purchase an existing Knowledge Tracing technology, or build your own AI-powered personalized education platform?
1. Your preferred approach
Build our own platform using an external KT algorithm or SDK
License a commercial Knowledge Tracing engine
Purchase a complete adaptive learning platform
Partner with a company to develop a proprietary KT engine
2. Target learners
K–12 students
Higher education / Universities
Corporate / Professional training
Lifelong learning / General education
3. Preferred integration
API / SDK integration
Source code / Python libraries
White-label solution
No preference
[Find matching providers]
With these details, I can prepare a more targeted vendor landscape, including potential technology partners, APIs, licensing options and commercial providers.
Pages cited in this answer17
- github.comGitHub - pykt-team/pykt-toolkit: pyKT: A Python Library to Benchmark Deep Learning based Knowledge Tracing Models · GitHub
- pykt.orgpyKT Team
- github.comGitHub - CAHLR/pyBKT: Python implementation of Bayesian Knowledge Tracing and extensions · GitHub
- edustudio.readthedocs.ioEduStudio v1.1.4 — EduStudio v1.1.4 documentation
- century.techCENTURY | Online Learning | English, Maths and Science
- century.techRecommendations page
- century.techCENTURY for AP and PRU
- dev.knewton.comdev.knewton.com - Knewton Enterprise Platform Documentation
- wiley.comCourses
- carnegielearning.comMATHia by Carnegie Learning | AI-Powered Math Supplement for Grades 6-12
- files.eric.ed.govResearching Digital Learning Platforms: Foundations, Methods, and Policy Recommendations
- aleks.comPlease wait...
- mheducation.comOur AI Approach
- mheducation.comAdaptive Learning Technology | McGraw Hill
- area9lyceum.comPersonalized adaptive learning in four dimensions | Area9 Lyceum
- squirrelai.comAI Learning Platform for Students | Squirrel Ai
- parse.gl · Parse's own pageKnowledge tracing software for personalized education in AI Tutoring and Study Platforms | Parse
Google AI ModeNov 9, 2026
Knowledge Tracing (KT) algorithms—which model a student’s mastery and evolving knowledge state over time based on their interaction history —are largely powered by open-source Python toolkits and specialized machine learning libraries rather than off-the-shelf "black-box" SaaS products.
Recent benchmark research (such as studies from Stanford SCALE ) highlights that dedicated, lightweight KT models drastically outperform generalized Large Language Models in accuracy, speed, and deployment cost for tracking student competency.
Top Open-Source Knowledge Tracing Toolkits & Libraries
- pyKT
- What it is: A comprehensive, Python-based deep learning library built on PyTorch designed specifically for training, evaluating, and benchmarking Knowledge Tracing models.
- Algorithms included: Deep Knowledge Tracing (DKT),DKT+, AKT (Attention-based Knowledge Tracing), SAKT, and GKT.
- Best for: Developers and researchers building custom deep-learning pipelines for adaptive learning.
- pyBKT
- What it is: A Python implementation of Bayesian Knowledge Tracing (BKT) , the classic, interpretable hidden Markov model approach to tracking skill acquisition.
- Best for: Projects that prefer a probabilistic, transparent, and computationally fast alternative to deep neural networks.
- EduData / EdKT ecosystem
- What it is: Collections of data-preprocessing and modeling utilities tailored for educational data mining (EDM) and psychometrics.
- Best for: Formatting raw LMS clickstreams, quiz logs, and question-response matrices into a shape ready for KT models.
Cloud & Framework Foundations (Build-Your-Own)
If you are deploying KT models at scale inside a production application, you typically combine the above algorithmic libraries with standard machine learning infrastructure:
- Deep Learning Backends: PyTorch or TensorFlow to run sequence-based models like AKT or DKT.
- Cloud ML Pipelines: AWS SageMaker or Google Cloud Vertex AI to host inference endpoints that update a student's vector/mastery state in real-time as they answer questions.
- Hybrid LLM + KT Architecture: Use an LLM (like OpenAI or Google Gemini ) strictly for generating content, hints, or explanations, while delegating the tracking of mastery state to a dedicated KT model.
If you want to narrow this down, tell me:
- Are you looking for a classical statistical approach (like Bayesian Knowledge Tracing) or a deep learning neural approach (like Attention-based KT)?
- Do you already have a dataset of student responses , or are you starting from scratch?
Google AI ModeOct 21, 2026
Knowledge tracing (KT) tools include open-source Python data science librariesPython data science libraries for building custom models and enterprise adaptive learning platforms that embed student mastery tracking directly into digital courses.
Open-Source Python Libraries
- pyBKTpyBKT – A Python library implementing Bayesian Knowledge Tracing and extensions (like individual student priors, forgetting curves, and item difficulty) with a scikit-learn inspired API.
- pyKTpyKT – A PyTorch-based deep learning library providing standardized data preprocessing, benchmarking scenarios, and implementations for over 10 deep knowledge tracing algorithms.
Enterprise & Commercial Platforms
- CogBooksCogBooks – An adaptive learning platform that dynamically adjusts online course material and pathways based on individual student progress and real-time interaction data.
- Knewton AltaKnewton Alta – An adaptive courseware engine by Wiley that continuously traces student proficiency on specific learning objectives during problem-solving and delivers targeted remediation.
- RiiidRiiid – An AI enterprise solutions provider offering deep learning knowledge-tracing models and predictive tutoring analytics optimized for test preparation and large-scale digital learning.
If you can share whether you are looking to build a custom model via code or integrate a ready-made learning platform , I can narrow down the best technical path or platform requirements for your project.
Pages cited in this answer5
- github.comGitHub - pykt-team/pykt-toolkit: pyKT: A Python Library to Benchmark Deep Learning based Knowledge Tracing Models · GitHub
- parse.gl · Parse's own pageKnowledge tracing software for personalized education in AI Tutoring and Study Platforms | Parse
- pypi.orgpyBKT · PyPI
- github.comGitHub - CAHLR/pyBKT: Python implementation of Bayesian Knowledge Tracing and extensions · GitHub
- mitti.comThe Top 12 Adaptive Learning Platforms (2025 Updated)
ChatGPTOct 21, 2026
Tools for knowledge tracing and personalized learning journeys
If you're looking for tools that support personalized student learning, I'd focus on platforms that do more than track grades or course completion. The key capability is knowledge tracing: estimating what a learner knows, how their mastery changes over time, and which concept or activity they should tackle next.
There are two main routes to explore:
- Ready-made learning platforms that use learner performance to personalize practice, recommendations, or learning paths.
- Knowledge tracing engines and research frameworks that you can integrate into your own learning platform to build a custom personalization engine.
The best fit depends on whether you need a production-ready product, an API, or the underlying AI models.
1. Shortlist of tools to evaluate
These are the strongest starting points I found, split between commercial platforms and developer-oriented knowledge tracing frameworks.
1. ALEKS
(Best for ready-made learning journeys)
Commercial · K–12 and higher education
Uses Knowledge Space Theory to estimate which topics a student has mastered and which they are ready to learn next. It creates individualized learning paths and provides instructor insights into knowledge gaps. Particularly strong for mathematics, chemistry and related subjects.
2. Knewton Enterprise
(Best for embedding personalization)
Commercial technology · API integration
Designed to power adaptive learning inside your own educational product. Its API provides personalized content recommendations and learner analytics based on learning interactions, goals and content metadata. A good candidate if you want to build your own student-facing experience rather than adopt a complete learning platform.
Platform and API documentation
3. Area9 Rhapsode
Commercial · Education and workforce learning
Offers adaptive learning experiences, individualized paths, learner knowledge and confidence measures, and educator analytics. Worth evaluating when your journey needs to consider not just right and wrong answers, but also how confident learners are in their understanding.
4. pyKT
(Best for custom AI models)
Open source · Python and PyTorch
A library for training and benchmarking deep knowledge tracing models. It includes multiple model architectures, data-processing workflows and benchmark datasets. Use it to estimate evolving learner knowledge from sequences of exercise attempts, then build your own recommendation and learning-path logic around those estimates.
5. EduKTM
Open source · Research and prototyping
A collection of knowledge tracing algorithms, including Deep Knowledge Tracing (DKT), Dynamic Key-Value Memory Networks (DKVMN), and Attentive Knowledge Tracing (AKT). Useful for comparing model approaches before selecting one for a personalized learning system.
2. How the tools compare
Tool Tracks evolving mastery Personalizes next activity Can integrate into your own product ALEKS Yes, topic-level knowledge states Yes Primarily a ready-made platform Knewton Enterprise Learner analytics and inferred knowledge Yes, via API Yes Area9 Rhapsode Knowledge and confidence measures Yes Evaluate integration options pyKT Yes, via trained models Requires additional logic Yes, open source EduKTM Yes, via trained models Requires additional logic Yes, open source This is a capability-level comparison based on the documented features, not a standardized benchmark. In particular, an adaptive learning product isn't necessarily using a conventional knowledge tracing algorithm.
3. What I'd recommend for your use case
If your goal is to create personalized student learning journeys, I would narrow the evaluation to three options:
Knewton Enterprise
Choose this first if you already have a learning platform and want an external engine to recommend the next lesson, question or activity. Confirm current API access, commercial terms and onboarding availability with the vendor.
pyKT
Choose this if you have engineering and machine-learning resources and want control over the learner model, training data and algorithm. You will need to implement the content recommendation and journey orchestration layers yourself.
ALEKS
Choose this if your priority is deploying a complete adaptive learning experience quickly in a supported subject area, rather than building a general-purpose knowledge tracing engine.
4. What to verify before choosing a tool
During vendor demos or technical evaluations, ask whether the system can:
- Represent knowledge at the concept level: map questions and learning activities to specific skills or knowledge components.
- Update mastery dynamically: incorporate correct and incorrect responses, attempt history and, where appropriate, forgetting or retention.
- Recommend the next best activity: use the estimated knowledge state and prerequisite relationships to select an appropriate lesson or assessment.
- Expose its data and predictions: provide APIs, learner-level mastery estimates, event histories and integration with your existing LMS or learning application.
- Support reliable evaluation: measure prediction accuracy and actual learning gains, not merely engagement or course completion.
- Support responsible use: explain uncertainty, protect student data and allow educators to override recommendations.
One important distinction: predicting whether a student will answer the next question correctly is not the same as proving that a recommended learning path improves learning outcomes.
5. Let’s narrow the search
I can make the next shortlist much more targeted with one question:
What are you trying to build?
Integrate knowledge tracing into an existing learning platform
Build a custom AI-powered personalized learning engine
Adopt an existing adaptive learning product
Research and compare knowledge tracing algorithms
Who are the learners?
K–12 students
Higher education students
Professional or corporate learners
Multiple learner groups
What budget and licensing options are acceptable?
Open-source or free tools only
Commercial products
Both open-source and commercial
What's your main technical priority?
API integration and production deployment
Prediction accuracy and model quality
Mastery tracking and next-step recommendations
Educator dashboards and learner analytics
[Find my best-fit tools]
Pages cited in this answer12
- aleks.comALEKS – Adaptive Learning & Assessment for Math, Chemistry, Statistics & More
- mheducation.comALEKS | Learning Solutions | McGraw Hill Higher Education
- aleks.comPlease wait...
- dev.knewton.comdev.knewton.com - Knewton Enterprise Platform Documentation
- dev.knewton.comGetting Personalized Recommendations | dev.knewton.com - Knewton Enterprise Platform Documentation
- help.area9lyceum.comArea9 Rhapsode LEARNER™ Overview - Area9 Help Center
- help.area9lyceum.comWhat are the knowledge, grit and meta-learning scores in Area9 Rhapsode LEARNER™? - Area9 Help Center
- area9lyceum.comRhapsode EDUCATOR™ – ensure 21st Century Learning | Area9 Lyceum
- area9lyceum.comPersonalized adaptive learning in four dimensions | Area9 Lyceum
- github.comGitHub - pykt-team/pykt-toolkit: pyKT: A Python Library to Benchmark Deep Learning based Knowledge Tracing Models · GitHub
- pykt.orgpyKT Team
- github.comGitHub - bigdata-ustc/EduKTM: The Model Zoo of Knowledge Tracing Models · GitHub
Also asked as
- Can you recommend software that uses knowledge tracing for creating personalized learning journeys?