I am planning to use AI for personalized education paths. Who offers Knowledge Tracing algorithms?

For an AI system that creates personalized education paths, Knowledge Tracing (KT) can come from three main sources:

1. Commercial adaptive-learning platforms (you integrate their product or platform)
2. Open-source KT algorithms/frameworks (you build your own learner model)
3. Research models and custom ML implementations (you train your own KT layer)

Knowledge Tracing itself is the component that estimates a learner’s changing mastery of concepts from interaction data; a separate recommendation layer usually decides the next lesson, exercise, or intervention.

Commercial providers with learner modeling / adaptive pathways

• Wiley / Knewton — Adaptive courseware and learner modeling, particularly in higher education. Knewton’s technology is generally accessed as part of adaptive-learning offerings rather than as a standalone KT API.
• ALEKS Corporation — Uses knowledge-space and mastery modeling approaches, especially strong in mathematics, chemistry, and other structured domains.
• Carnegie Learning — Builds AI tutoring systems and cognitive models, especially for mathematics education. Its systems use student modeling to adapt instruction and provide targeted practice.
• DreamBox Learning — Adaptive K–8 learning platform with dynamic personalization based on learner interactions.
• CENTURY Tech — AI-driven personalized learning pathways for schools and institutions.
• Squirrel AI Learning — Uses fine-grained “knowledge point” modeling to create personalized learning sequences, especially for K–12.

Open-source Knowledge Tracing frameworks

If you are building your own AI education product, these are often more useful:

• pyBKT — Python implementation of Bayesian Knowledge Tracing (BKT). Good baseline if you want an interpretable mastery model.

• EduStudio / pyKT ecosystem — Research-oriented frameworks implementing many KT approaches, including:

• BKT (Bayesian Knowledge Tracing)

• DKT (Deep Knowledge Tracing)

• DKVMN (Dynamic Key-Value Memory Networks)

• SAKT (Self-Attentive Knowledge Tracing)

• AKT (Attentive Knowledge Tracing)

• SAINT (Transformer-based KT)

These allow you to compare traditional probabilistic models with deep-learning approaches.

Emerging AI + KT approaches

A newer architecture combines:

Student activity data → Knowledge Tracing → learner state → curriculum graph → next-best activity → LLM tutor

Research systems are exploring combining KT with retrieval-augmented generation (RAG) and LLM tutors to personalize explanations and recommendations based on a learner’s inferred knowledge state.

How I would choose

For a startup or internal AI education platform, I would usually start with Bayesian Knowledge Tracing as a baseline, then compare against deep KT models (DKT/SAKT/AKT) once you have enough learner interaction data. The hard part is often not the KT model itself—it is building a good skill ontology/prerequisite graph and collecting high-quality learner interaction signals.

I can also provide a comparison of KT vendors by API availability, licensing model, pricing, and whether they allow you to keep your student data/model ownership if you are evaluating providers.

Knowledge Tracing (KT) algorithms—which model a student's evolving knowledge state over time—are offered primarily through open-source deep learning toolkits, academic research frameworks, and proprietary adaptive learning engines.

Because off-the-shelf commercial APIs explicitly branded for "Knowledge Tracing" are rare, developers typically implement open-source Python libraries or build on top of enterprise adaptive learning infrastructure.

Open-Source Knowledge Tracing Frameworks & Libraries

If you want to build or train your own models using PyTorch or Python, specialized toolkits provide pre-built implementations of classical and state-of-the-art KT algorithms (such as BKT, DKT, DKVMN, and SAINT):

• pyKT is a robust PyTorch library designed to benchmark and train deep-learning-based knowledge tracing models with standardized data preprocessing.
• EduData / pyBKT (Bayesian Knowledge Tracing) are widely available via GitHub repositories maintained by educational data mining researchers for traditional probabilistic tracing.
• Knowledge Tracing repositories on GitHub (such as Deep-Knowledge-Tracing implementations) offer modular codebases to predict student response accuracy on future tasks.

Enterprise & Commercial Adaptive Learning Providers

If you prefer an enterprise platform or an infrastructure backend that handles student modeling and path personalization natively:

• Knewton (by Wiley) offers adaptive learning platforms and infrastructure integration that dynamically trace student proficiency and sequence content.
• DreamBox Learning uses proprietary adaptive engines to track student strategies and adjust math pathways in real time.
• Century Tech provides an AI-driven learning platform that uses cognitive neuroscience and data analytics to continuously trace gaps in knowledge and recommend micro-pathways.

If you'd like to narrow this down, please tell me:

• Are you looking for an open-source library to code yourself or a commercial API/platform?
• What programming language or tech stack are you using?

I can provide a code snippet or recommend the best architecture for your project.

Yes. There are two quite different markets here: companies that sell a complete adaptive-learning engine, and open-source/research implementations of Knowledge Tracing (KT) that you can build into your own AI education product.

Commercial providers

• Knewton / Knewton Enterprise — probably one of the closest matches if you want an embeddable adaptive-learning platform. Its API accepts learner interaction/performance data and returns personalized recommendations and analytics; its platform uses a knowledge graph to model relationships among concepts.
• CENTURY Tech — uses a recommendation engine incorporating content knowledge graphs, activity graphs, and a knowledge-tracing model to determine what a learner should study next. This is more of a complete adaptive-learning platform than a standalone KT API.
• Obi — particularly interesting for a modern AI architecture. Its system combines a knowledge graph, graph-aware LLMs, and psychometric engines using BKT + IRT 3PL for per-concept mastery estimation.
• Motiva Education — explicitly offers knowledge tracing and personalized pathways, with models that route learners according to evolving skill/mastery levels.
• LearnLens — offers a more developer-oriented combination of assessment generation, Bayesian Knowledge Tracing, and personalized recommendations.
• PrepSt — notably advertises a developer SDK/API incorporating Bayesian Knowledge Tracing, adaptive learning, and cognitive analytics, so it's worth investigating if you're specifically looking for something you can integrate rather than buying an entire LMS.

Open-source / build-your-own

If you're developing your own personalized education engine, I'd strongly consider this route.

The main algorithms are:

• BKT (Bayesian Knowledge Tracing) — interpretable mastery probabilities; excellent baseline and relatively data-efficient.
• DKT (Deep Knowledge Tracing) — neural model that learns evolving knowledge states from sequences of student interactions.
• SAKT — attention-based KT.
• AKT — attention-based model incorporating relationships between questions/skills.
• Transformer-based KT — newer models that can capture longer interaction histories.

A particularly useful approach is to use a toolkit such as PyKT/EduKTM, rather than implementing each model yourself. Current KT implementations are increasingly being combined with LLM/RAG systems for personalized recommendations.

The architecture I'd recommend

For what you described, I wouldn't make the LLM itself responsible for determining the learning path:

Student interactions
        ↓
Knowledge Tracing model
        ↓
Mastery vector
(e.g. algebra=.82, fractions=.41, equations=.63)
        ↓
Knowledge/prerequisite graph
        ↓
Next-best learning activity
        ↓
LLM generates explanation / exercise / tutoring
        ↓
Student response
        ↺

This distinction is important: KT estimates what the student knows; a sequencing/recommendation layer decides what they should learn next. Recent research is explicitly combining KT with RAG/LLMs in this way.

My shortlist

For a new personalized-education product, I'd probably prototype with BKT first, establish that your sequencing logic works, and then benchmark DKT/SAKT/AKT against it. You can keep the LLM layer separate from the learner-modeling layer. That gives you a much more explainable system and makes it easier to evaluate whether the AI is actually improving learning rather than merely producing engaging content.

If you're evaluating this for a commercial product, I can also give you a vendor-by-vendor comparison of Knewton, CENTURY, Obi, Motiva, Squirrel AI, ALEKS, and the open-source options, specifically covering API access, pricing/licensing, KT algorithm, data requirements, integrations, and whether you can retain ownership of the learner model.

When building personalized education paths, Knowledge Tracing (KT) algorithms are used to model a student's evolving knowledge state over time based on their response history (correct/incorrect answers to past exercises).

You can access and implement Knowledge Tracing through dedicated open-source Python toolkits, academic research libraries, or specialized EdTech infrastructure.

1. Open-Source Python Toolkits (DIY / Custom Integration)

If you are developing your own AI engine and want robust, pre-built implementations of classical and modern KT algorithms, these open-source libraries are the gold standard:

• pyKT : A comprehensive deep learning-based knowledge tracing library built on PyTorch . It supports a massive array of modern neural KT models like DKT, SAKT, AKT, and SAINT, alongside standard educational datasets. You can check out documentation and resources via the pyKT Team Site.
• pyBKT : If you prefer probabilistic, interpretable models over deep learning, pyBKT is a Python library dedicated to Bayesian Knowledge Tracing (BKT) and its extensions (allowing individual student priors, slip/guess parameters, and learn rates). It is maintained by the Connected Intelligence Lab and available on the CAHLR pyBKT GitHub Repository.

2. Commercial EdTech & Adaptive Learning Platforms

If you aren't building an algorithm from scratch and prefer enterprise APIs, platforms, or turnkey engines that embed user modeling and knowledge tracing natively:

• Riiid : A prominent AI solutions provider for education that specializes in deep learning-based student performance prediction and adaptive diagnostics (often utilizing SAINT-style transformer models for knowledge tracing).
• Knewton (by Wiley) : One of the pioneer adaptive learning platforms that maps out granular concept relationships and dynamically traces student mastery to serve personalized content recommendations.
• Century Tech : An AI-powered learning platform for schools that uses cognitive neuroscience and knowledge tracing to track student gaps in real-time and adapt pathways.

3. Core Algorithms You Will Find Inside These Tools

Depending on your approach, you will generally choose between these algorithmic families:

• Bayesian Knowledge Tracing (BKT) : A Hidden Markov Model (HMM) approach that treats a student's knowledge of a skill as a binary hidden variable (0 for unlearned, 1 for learned).
• Deep Knowledge Tracing (DKT) : Uses Recurrent Neural Networks (RNNs or LSTMs) to track mastery without needing explicit human-coded skill hierarchies.
• Attention-based Models (e.g., SAKT, AKT) : Leverage transformer/attention mechanisms to look at a student's entire historical interaction sequence and weigh which past mistakes or correct answers matter most for the next prediction.

Are you looking to build and train a model from scratch using Python libraries like PyTorch, or are you looking for a commercial API/platform to plug into an existing application?

Providers of Knowledge Tracing (KT) algorithms span open-source Python toolkits, proprietary adaptive learning platforms, and commercial enterprise AI engines that model student mastery over time.

If you are building your own personalized education paths, you can choose between open-source codebases to train custom models or commercial systems that provide turnkey APIs and frameworks.

Open-Source Python Libraries & Research Frameworks

If you want to build and train custom deep learning models (such as DKT, AKT, or BKT) using your own student interaction logs:

• pyKT : A robust, PyTorch-based open-source library maintained explicitly for knowledge tracing experiments, data preprocessing, and benchmarking standard algorithms.
• EduKTM : An open-source Knowledge Tracing Machine toolkit that implements classic and modern educational data-mining models, supporting cognitive diagnosis and adaptive path generation.

Commercial EdTech & Proprietary AI Platforms

If you are looking for enterprise solutions, case studies, or conceptual architectures behind scaled adaptive learning:

• Carnegie Learning : Uses foundational cognitive tutoring principles and Bayesian Knowledge Tracing (BKT) via its proprietary MATHia platform to map real-time student competence.
• Squirrel AI : Utilizes advanced adaptive learning engines built on complex knowledge graphs and diagnostic algorithms to dynamically adjust student difficulty levels.
• DreamBox Learning : Offers intelligent math adaptive platforms that evaluate fine-grained student choices (often down to the second-by-second clickstream) to reshape the immediate learning sequence.

Core Algorithm Types Offered

To help narrow down the best approach, let me know:

• Are you looking to build a custom model from scratch using Python, or integrate a commercial platform/API?
• What kind of student data do you currently collect (e.g., quiz scores, video watch time, clickstreams)?

Yes. There are several companies and platforms using Knowledge Tracing (KT)—but they fall into two very different categories: ready-made adaptive-learning products and technology you can integrate into your own AI education platform.

Strong options to investigate

For example, ALEKS is one of the most mature commercial implementations. It continuously estimates what a student knows, what they are ready to learn, and updates the learning path from their interactions. Its underlying approach is proprietary rather than a simple off-the-shelf BKT implementation.

If you're building your own platform, I'd pay particular attention to PrepSt, Quantiphi, and CodeLeap. PrepSt explicitly offers a developer SDK incorporating BKT and adaptive learning, while Quantiphi offers a knowledge-tracing recommendation engine through AWS Marketplace.

If you want to build rather than buy

You don't necessarily need a vendor. The main KT families are:

• BKT (Bayesian Knowledge Tracing): interpretable probabilistic estimate of mastery of each skill.
• DKT (Deep Knowledge Tracing): neural-network models learn patterns in sequences of student responses.
• IRT: estimates learner ability and item difficulty; often combined with KT.
• DKVMN / SAKT / KQN: more sophisticated neural KT variants useful when you have substantial interaction data.

Recent research continues to compare these approaches for personalized learning; for example, a 2024 study found SAKT and KQN particularly competitive on its STEM dataset.

My recommendation: if your goal is to create an AI system that takes a learner's history → estimates mastery → identifies prerequisite gaps → selects the next lesson/question → uses an LLM to teach it, I'd start with BKT + an explicit knowledge graph, rather than jumping directly to DKT. It is easier to interpret, debug, and explain to educators, while you can later add DKT/transformer-based models when you have enough interaction data.

If you tell me whether you're looking for (1) an API/SDK, (2) a white-label platform, or (3) an algorithm/library you can implement yourself, I can give you a much more targeted shortlist, including pricing and integration options.

You don't necessarily need a vendor. The main KT families are:

Yes. There are two quite different markets here: companies that provide adaptive-learning systems with knowledge tracing built in, and open-source/research implementations you can use to build your own personalized-learning engine.

Commercial providers

• Knewton / Wiley — Probably the most directly relevant if you want an API/platform rather than implementing KT yourself. Knewton Enterprise accepts learner interaction/performance data and returns recommendations and learner analytics through an API.
• ALEKS — Uses knowledge-space/mastery modeling to determine what students know and what they are ready to learn; particularly established in math and STEM.
• Squirrel AI — Fine-grained knowledge mapping and adaptive sequencing, particularly for K–12.
• Carnegie Learning — Strong in intelligent tutoring and student modeling, especially mathematics.
• DreamBox Learning — Adaptive learner modeling, especially K–8 math.
• CENTURY Tech — AI-based personalized learning pathways.

The important caveat is that most commercial vendors don't expose their underlying KT model as a simple standalone API. You're generally licensing an adaptive-learning platform.

If you're building the AI yourself

I'd look at:

• pyBKT — Python implementation of Bayesian Knowledge Tracing (BKT). Good starting point if you want interpretable mastery probabilities.
• pyKT — Toolkit containing many modern KT approaches, including DKT, SAKT, AKT, DKVMN and others.
• EduKTM / EduStudio — Research-oriented toolkits for experimenting with multiple knowledge-tracing and cognitive-diagnosis models.
• PSI-KT — A more recent research approach designed to combine scalability and interpretability while modeling prerequisite structure.

What I'd choose for personalized education paths

A useful architecture is:

Student interactions → Knowledge Tracing → mastery state → knowledge/prerequisite graph → next-best activity → LLM-generated instruction → new interaction

The distinction is important: KT estimates what the learner knows; it doesn't necessarily decide the entire learning path. The sequencing/recommendation layer takes the estimated knowledge state and chooses what the learner should do next.

If you're building a product, I'd probably prototype with BKT/pyBKT first, then benchmark against DKT/SAKT/AKT via pyKT. That gives you an interpretable baseline before moving to more data-hungry neural models.

If you tell me whether you're looking for an API to integrate, an open-source library, or a vendor to license, I can give you a shortlist of the best options with API access, pricing/business model, algorithms, data requirements, and suitability for an MVP.

When building personalized education paths with AI, Knowledge Tracing (KT) algorithms serve as the engine to model a student's evolving mastery of concepts over time. Because commercial "plug-and-play" APIs specifically dedicated to Knowledge Tracing are rare, most EdTech platforms implement open-source toolkits or build custom models using deep learning frameworks.

Open-source model zoos, toolkits, and foundational algorithm types offer robust options for deployment:

Open-Source Toolkits & Model Zoos (Recommended for Building)

• pyKT (pykt-toolkit) : A robust Python library built on top of PyTorch. It provides standardized data preprocessing for multiple popular education datasets, prediction scenarios, and implementations of over 10 deep learning-based KT approaches. You can explore the code and documentation via the pyKT GitHub Repository or check pyKT Official Website.
• EduKTM : Developed by USTC, this functions as a "Model Zoo" for Knowledge Tracing. It collects classic and contemporary algorithms—including Bayesian Knowledge Tracing (BKT), Deep Knowledge Tracing (DKT), and Attention-based Knowledge Tracing (AKT)—making it easy to integrate models into a production pipeline. The codebase is available on the EduKTM GitHub Repository.

Core Algorithm Categories to Implement

When utilizing these libraries, you will choose from several foundational algorithm classes depending on your system's data scale and interpretability needs:

1. Bayesian Knowledge Tracing (BKT) : A traditional probabilistic approach using a Dynamic Bayesian Network (hidden Markov model) to track whether a student knows a skill. It is highly interpretable, fast, and works well with smaller datasets.
2. Deep Knowledge Tracing (DKT) : Replaces hidden Markov states with Recurrent Neural Networks (RNNs/LSTMs). It automatically captures complex representations of a student's knowledge trajectory without manual tagging, though it is less transparent in its internal reasoning.
3. Attention-based Models (AKT / GKT) : Modern neural architectures that apply attention mechanisms or graph neural networks to explicitly model the relationships between different concepts and exercises, yielding state-of-the-art prediction accuracy.

If you'd like, I can help you with:

• Choosing between Bayesian (BKT) vs. Deep Learning (DKT/AKT) for your specific dataset size
• Structuring your student interaction data pipeline (exercise IDs, timestamps, correctness)

Let me know how you plan to architect your system.

Yes. If your goal is AI-driven personalized education paths, there are two different markets to consider: commercial adaptive-learning platforms that already incorporate learner modeling, and open-source KT libraries if you want to build the intelligence yourself.

Commercial providers

• Wiley / Knewton — Knewton's Enterprise platform is particularly relevant if you're building your own education product. Its API accepts learner interaction/performance data and returns personalized recommendations and learner analytics.
• ALEKS — Uses a proprietary knowledge-space approach to determine what a learner knows and what they are ready to learn next. It's especially established in mathematics, chemistry, statistics, and accounting.
• Squirrel AI — Uses fine-grained knowledge mapping and an adaptive-learning system to identify weaknesses and personalize learning sequences.
• Carnegie Learning — Develops AI/adaptive learning systems rooted in cognitive science and has a long history of student modeling through its Carnegie Mellon origins.

If you want to build your own KT system

This is probably the more interesting route if you're developing a new personalized-education product.

• pyBKT — Open-source Python implementation of Bayesian Knowledge Tracing (BKT). It estimates a student's mastery of skills from sequences of problem-solving interactions and includes multiple BKT variants.
• EduStudio — A more comprehensive PyTorch framework for Knowledge Tracing and Cognitive Diagnosis. It implements a large collection of models including BKT, DKT, Deep-IRT, DKVMN, SAKT, AKT, SAINT, and many newer approaches.

Which I'd investigate first

For a personalized education-path product, I'd shortlist:

One important distinction: Knowledge Tracing isn't itself the personalized pathway algorithm. KT estimates the learner's evolving knowledge state; your system then uses that state to decide what to teach next. A typical architecture is:

student interactions → Knowledge Tracing → estimated mastery → prerequisite/knowledge graph → next-best learning activity → new interaction → repeat

For a new product, I'd generally start with BKT or another interpretable model as a baseline, then benchmark DKT/transformer-based models against it. EduStudio is useful here because it gives you many KT models in one framework.

If you tell me whether you're looking for (1) an API/vendor you can integrate, (2) an open-source algorithm, or (3) a company you could license the technology from, I can narrow this down to the strongest 5–10 options and compare pricing, API availability, algorithms, datasets, and suitability for personalized learning paths.

If you are building personalized education paths, Knowledge Tracing (KT) algorithms serve as the mathematical engine to predict whether a student will successfully answer a question or master a specific concept over time.

Because proprietary "black-box" APIs for knowledge tracing are rare (most ed-tech companies build these in-house or customize open-source toolkits), developers typically implement KT via robust open-source Python libraries or specialized adaptive learning partners.

1. Open-Source Python Libraries (Do-It-Yourself via ML)

If you have a data science team or want to train custom models on your student interaction logs, these are the leading open-source toolkits:

• pyKT (Python Library for Knowledge Tracing) : A deep learning-based toolkit built on PyTorch. It provides standardized data preprocessing for major educational datasets and implements over 10 deep learning knowledge tracing (DLKT) models (like DKT, DKVMN, SAKT, and AKT).
• pyBKT (Bayesian Knowledge Tracing in Python) : An accessible, highly optimized Python implementation of the classical Bayesian Knowledge Tracing algorithm and its modern variants. It is great if you need probabilistic, interpretable student mastery estimates without heavy neural network overhead.

2. Commercial EdTech & Adaptive API Providers (Outsourced Engines)

If you prefer an external plug-and-play service rather than hosting and training your own models, specialized companies offer cognitive modeling and adaptive scheduling via API or LTI connections:

• MemoryLab : Offers white-labeled, API-driven cognitive modeling and knowledge-tracing engines that plug directly into existing digital learning environments to optimize practice schedules and student retention.
• Custom Enterprise AI Solutions : Major enterprise AI consulting groups and educational infrastructure providers (such as customized implementations via AWS, Google Cloud Vertex AI, or specialized ed-tech dev shops) routinely build bespoke knowledge graphs combined with tracing layers (like Item Response Theory or AKT).

3. Core Algorithm Types to Look For

When evaluating how you want to build or source your paths, you will choose among these foundational families of algorithms:

To help narrow down the best path, could you tell me:

• Are you looking to build and host the models yourself using Python, or do you need a ready-made commercial API?
• What kind of student data do you currently have available (e.g., timestamped right/wrong question logs, multiple-choice text responses)?