NewFive classroom tools co-designed with K–12 teachers in ExLAIM 2026

AI help that stays grounded in your course.

MerryQuery is an AI learning platform from NC State. Students get step-by-step answers that cite their instructor's own materials. Teachers decide how much the AI reveals, see where students struggle, and plan, practice, and give feedback from one course toolkit.

Open MerryQuery Explore the platform

Best Technical Demonstration Award, AAAI 2025

MerryQuery teacher workspace: the course sidebar and the course toolkit, grouped into planning, material creation, grading and feedback, and classroom activity.

MerryQuery student chat: a five-step explanation of a MATLAB for loop with a practice prompt and a citation to D.1.1-forLoops, tagged Hint Only. — Student viewCited · Step-by-step · Course-aware

students supported across three semesters: 750+
classroom tools co-designed with nine K–12 teachers: 5
AAAI Best Technical Demonstration Award: 2025

The platform

Who it's for

For students
Help that stays tied to what was actually taught.
- Step-by-step answers that cite the course materials
- Study Coach, Notes, Flashcards, and Practice Sets
- A teacher or classmate when a person would help more
See how it works
For teachers
A whole course toolkit, with you in charge of the AI.
- Nine tools for planning, materials, feedback, and class activity
- Five levels of AI help, set for each course assistant
- Analytics on performance, engagement, and support
See how it works

For students

MerryQuery answers from the materials an instructor uploads, so help matches what was actually taught: step-by-step, with a citation to the source and a chance to practice.

Study tools that come with the course

Beyond chat, students get a Study Coach that helps plan their next step, Notes from their conversations, quick Flashcards, and the Practice Sets their teacher shares. When a person would help more, MerryQuery offers to bring in a teacher or classmate.

Every answer is grounded in instructor-uploaded materials and cites its source
Study Coach, Notes, Flashcards, and Practice Sets in one place
Team Mode connects students with a teacher or classmate without sharing their full chat

MerryQuery student Tools page: Notes, Flashcards, Practice Sets, and Study Coach.

For teachers

Plan, create materials, give feedback, and run classroom activities from one course workspace, decide how much help the AI gives, and see where students need support.

Tools

Your course toolkit

Every course includes teacher tools that draw on the same course materials. Most were built or rebuilt in 2026 with K–12 teachers.

How they were co-designed

Planning

Prepare lessons and teaching support.

Instructor Chat
Plan lessons, prepare materials, and ask course-specific teaching questions. Highlight any AI answer for Teaching Guidance on how to explain it.
ExLAIM 2026: how Instructor Chat was co-designed
Lesson Planner
Design classroom-ready lesson sequences aligned to your course goals, with citations to your materials.
ExLAIM 2026: how Lesson Planner was co-designed

Material creation

Build course-ready learning materials.

Flashcard Generator
Create concise study sets from selected course files and key concepts, then export them.
ExLAIM 2026: how Flashcard Generator was co-designed
Practice Sets
Generate, publish, and manage practice questions, with live topic-level analytics.
ExLAIM 2026: how Practice Sets was co-designed
Quiz Generation
Generate a quiz from selected documents, then preview, edit, publish, and monitor attempts.
ExLAIM 2026: how Quiz Generation was co-designed

Grading & feedback

Review learning evidence and respond to student work.

Feedback Generator
Turn uploaded student work and rubric criteria into an editable written feedback draft.
ExLAIM 2026: how Feedback Generator was co-designed

Classroom activity

Share daily updates and coordinate live support.

Daily Activity
Post updates, questions, warm-ups, and exit tickets, and see who has read or responded.
ExLAIM 2026: how Daily Activity was co-designed
Collaborative Board
Run shared Miro boards from inside your course, with a board-aware AI Coach.
ExLAIM 2026: how Collaborative Board was co-designed
Team Mode
Bring teachers and classmates into the learning process when human support matters most.
ExLAIM 2026: how Team Mode was co-designed

Control

Set how much help the AI gives

Every answer follows one of five GenAI assistance levels, from hints only to full solutions. Teachers set each assistant's default level; students can move between levels as they work, and every question is logged with the level it used.

Level-aware scaffolding guides students without giving answers away
Full solutions are marked learning-only and need the student's confirmation
Teachers see the level behind each student question

MerryQuery chat with the Mode menu open, showing the five assistance levels: Hint Only, Review, Troubleshoot, Co-Builder, and Full Solutions. The answer above is tagged Hint Only.

Insight

See where students need help, while it matters

Course analytics draw on evidence the tools already collect, so teachers can review performance, participation, collaboration, and support needs in one place, and the Review Center holds the questions the AI should not answer alone.

Completion, scores, and topic mastery across practice sets and quizzes
Daily engagement and collaborative-board activity
Open help requests, repeated support topics, and response times

MerryQuery course analytics page with four areas: Assessment Performance, Daily Engagement, Collaborative Learning, and Support & Intervention.

Under the hood

Built into every course

Source citations

Every grounded answer names the course document it drew from, so students can verify it and instructors can trust it.

Multimodal course materials

Complex PDFs with code, tables, and images are processed with OCR and text mining, organized in folders, and indexed in the background.

Tone and answer format

Teachers can let students choose a friendly, formal, encouraging, or Socratic tone, and request formats such as lesson plans for their own prep.

TAs and roles

Invite TAs and graders with scoped, role-based permissions and a dedicated interface for their responsibilities.

Review Center

Look back at student conversations when it helps, and see unclear or flagged questions gathered in one place for review.

Supports research

An experiment framework supports A/B studies, assigning conditions by lab section.

Summer 2026 · NSF-funded

Teachers and developers, building together

In ExLAIM, nine K–12 teachers and undergraduate developers spent six weeks turning classroom concerns into working MerryQuery tools: Team Mode, Collaborative Boards, Daily Activity, a Feedback Generator, and Practice Sets & Quizzes.

active weeks: 6
K–12 teachers: 9
undergraduate developers: 7
classroom tools: 5

Inside ExLAIM

News

Latest from MerryQuery

EXLAIMJul 2026
Five new classroom tools
ExLAIM teams presented Team Mode, Collaborative Boards, Daily Activity, the Feedback Generator, and Practice Sets & Quizzes, each built inside MerryQuery with K–12 teachers.
PAPERJun 2026
Assignment Perturbation at AIED 2026
Sam Gilson's study on how adversarial assignment perturbation affects students' help-seeking, using MerryQuery, was presented at AIED 2026 with co-authors Benyamin Tabarsi and Tiffany Barnes.
EXLAIMJun 2026
ExLAIM kicks off
Seven undergraduate developers (two development leads and five developers) and nine K–12 teachers began a six-week co-design program on MerryQuery.
FEATURESpring 2026
GenAI Usage Scale
Replaced the Practice/Review binary with a 5-level GenAI assistance scale, giving instructors fine-grained control over AI help.
FEATURESpring 2026
Analytics Dashboard
Launched the Instructor Analytics Dashboard with engagement metrics, struggle detection, and topic performance charts.
PAPERFeb 2026
MerryQuery at SIGCSE 2026
A condensed version of MerryQuery's usage analysis, covering how students use and perceive the assistant, appeared at SIGCSE 2026.

Benyamin Tabarsi explains the MerryQuery poster and live demo to two visitors at AAAI 2025. — Benyamin Tabarsi presenting the MerryQuery demonstration at AAAI-2025. Photo: NC State Computer Science.

Recognition

Best Technical Demonstration, AAAI 2025

“By balancing AI assistance with pedagogical oversight, it empowers both students and educators.”
Dr. Tiffany Barnes, Distinguished Professor, NC State

Read the NC State story

Team

The people behind MerryQuery

MerryQuery is led by researchers at NC State and Kennesaw State and built with students, instructors, and K–12 teachers.

Research Lead

Benyamin Tabarsi

Benyamin Tabarsi is a Ph.D. candidate in Computer Science at NC State, advised by Tiffany Barnes and DK Xu. His research focuses on human–AI co-intelligence. He aims to build AI that people can steer, so it helps them do more while they keep learning and growing their own skills. He designed and researched MerryQuery, and leads ExLAIM, an NSF-funded program where K–12 teachers and undergraduate developers design classroom AI tools together. In his work, he builds systems, studies them in real classrooms, and designs alongside the people who use them.

Website

Principal Investigator

Dr. Tiffany Barnes

Tiffany Barnes is a Distinguished Professor of Computer Science at NC State, where she leads the Game2Learn Lab. Her NSF CAREER Award recognized her work using data and educational data mining to add intelligence to STEM learning environments. Her current projects include data-driven tutoring in STEM, teaching AI to middle schoolers, and helping K–12 teachers bring AI and computational thinking into their courses. She is a founding co-director of the STARS Computing Corps and an ACM Distinguished Member.

Website

Principal Investigator

Dr. Dongkuan (DK) Xu

Dongkuan (DK) Xu is an Assistant Professor of Computer Science at NC State, where he leads the Generative Intelligent Computing (GIC) Lab. His research studies learning in open-ended environments: how AI systems decide what is worth exploring and how they should change as experience accumulates, with a focus on efficient learning, reliable AI, and continual model evolution. His honors include the AAAI 2025 Best Demonstration Award and AAAI New Faculty Highlights.

Website

Co-PI, ExLAIM

Dr. Joey Huang

Joey Huang is an Assistant Professor of Learning, Design, and Technology in NC State's Department of Teacher Education and Learning Sciences and a Co-PI and faculty mentor of ExLAIM. Trained in educational psychology and computing, she studies computational thinking, AI-enabled educational technologies, and how learners co-design with AI in STEAM. She also brought MerryQuery into her graduate course, ECI 519.

Website

Co-PI, ExLAIM

Dr. Xiaoyi Tian

Xiaoyi Tian is an Assistant Professor in the Department of Information Technology at Kennesaw State University and a Co-PI and faculty mentor of ExLAIM. Her research brings together human-computer interaction, AI, and the learning sciences to build human-centered AI that supports equitable teaching and learning, from learning companions and teacher-facing tools to AI literacy in K–12 education. She was previously a Research Scientist at NC State.

Website