Diagnose misconceptions,
not mistakes.
SciSleuth runs short diagnostic assessments, detects the specific physics misconceptions behind each wrong answer, and maps them onto a live knowledge graph.
Student answer
“Heavier objects fall faster because gravity pulls more on them.”
SciSleuth diagnosis
Misconception detected: force confused with acceleration. Knowledge graph node Newton's 2nd law flagged unhealthy.
The problem
Grading marks the answer.
Nobody marks the thinking.
A wrong answer is just a symptom. Underneath it lives a misconception — a stubborn mental model that keeps producing wrong answers across topics. Standard grading never surfaces it, and teachers have no view of which ideas are actually broken.
1 score
Tests give a single number — no signal about which underlying concept failed.
0 maps
Teachers have no map of which concepts in a unit are healthy vs. broken.
Hours
Diagnosing a single student's misconceptions by hand takes a 1-on-1 conversation.
How it works
From a diagnostic quiz to a recovery mission.
Run a diagnostic
Students take a short diagnostic assessment designed to trigger specific misconceptions, not just measure correctness.
Detect & explain
Gemini analyzes each response, names the underlying misconception, and writes a plain-language explanation of the flawed reasoning.
Recover
SciSleuth generates a targeted AI recovery mission per student and updates their knowledge graph health score.
What ships today
Seven pieces,
one diagnostic loop.
Every feature listed here is implemented in the current SciSleuth build.
Diagnostic assessment
Short, targeted quizzes engineered to expose specific physics misconceptions rather than just measure correctness.
Misconception detection
Each response is classified against a misconception catalog so wrong answers are tagged with the idea behind them.
Knowledge graph
Concepts and their prerequisites are linked in a graph so a broken node visibly affects everything downstream of it.
Graph health scoring
Each node carries a health score that updates as the student answers, giving an at-a-glance view of conceptual mastery.
AI explanations
Gemini generates a plain-language explanation of the flawed reasoning behind every detected misconception.
AI recovery missions
Per-student missions are generated to directly target the misconceptions that scored lowest on the knowledge graph.
Teacher analytics
Teachers see which misconceptions are most common across the cohort and which knowledge graph nodes need intervention.
Stop grading mistakes.
Start diagnosing minds.
Try the live SciSleuth diagnostic — take a quiz, see your knowledge graph update, and get an AI recovery mission.