Conker

Accepts educational content (text, documents, or topic descriptions) and uses LLM-based generation to automatically create multiple-choice, short-answer, and fill-in-the-blank questions with corresponding answer keys. The system likely employs prompt engineering to control question difficulty, cognitive level (Bloom's taxonomy alignment), and question type distribution, reducing manual authoring time from hours to minutes while maintaining pedagogical validity.

Provides educators with controls to specify question difficulty (basic, intermediate, advanced), cognitive complexity (recall, comprehension, application, analysis), and question type distribution before generation. The system maps these specifications to LLM prompt parameters and generation constraints, enabling creation of differentiated assessments for mixed-ability classrooms without generating separate quizzes manually.

Dynamically adjusts quiz difficulty or question selection based on student responses in real-time, presenting easier questions to struggling students and harder questions to high performers. The system uses item response theory (IRT) or Bayesian adaptive testing algorithms to estimate student ability and select next questions with optimal difficulty. Likely stores student ability estimates and question difficulty parameters in a database for ongoing calibration.

Automatically generates alternative text for images, optimizes question formatting for screen readers, ensures color contrast compliance, and produces adjustable text size variants during quiz creation. The system integrates accessibility checks into the generation pipeline (not as post-processing), producing WCAG 2.1 AA-compliant content by default. Likely uses accessibility metadata standards (ARIA labels, semantic HTML) and image description LLM models to generate contextually appropriate alt text.

Hosts quizzes on Conker's platform and collects student responses in real-time, tracking completion status, response timing, and answer correctness. The system provides educators with live dashboards showing class-wide performance metrics, individual student progress, and question-level analytics. Likely uses WebSocket or polling for real-time updates and stores response data in a relational database with indexing for fast analytics queries.

Allows educators to save generated or manually-created questions to a persistent question bank, organize questions by topic/standard/difficulty, and reuse questions across multiple quizzes. The system provides search and filtering capabilities (by keyword, difficulty, question type, learning objective) and likely uses tagging or metadata indexing to enable fast retrieval. Supports bulk operations (import/export, batch tagging) for managing large question libraries.

Analyzes quiz response data to generate reports showing class-wide performance trends, individual student mastery levels, question-level difficulty/discrimination metrics, and learning gap identification. The system calculates statistics (mean score, standard deviation, item difficulty, point-biserial correlation) and visualizes results in dashboards and exportable reports. Likely uses statistical analysis libraries and data aggregation queries to compute metrics from response logs.

Integrates with learning management systems (Canvas, Google Classroom, Blackboard, Schoology) via LTI (Learning Tools Interoperability) protocol or direct API connections, enabling educators to launch quizzes from within their LMS and automatically sync grades back to the gradebook. Supports SSO via OAuth 2.0 or SAML for seamless authentication without separate login. Likely uses LTI 1.3 standard for secure, standards-based integration.

Delivers quizzes via responsive web design that adapts to mobile devices (phones, tablets) and desktop browsers, with touch-optimized controls for mobile input. May include limited offline capability (caching quiz content locally) to allow students to complete quizzes without continuous internet connectivity, with response synchronization when connection is restored. Likely uses service workers or progressive web app (PWA) patterns for offline support.

Allows educators to import quizzes from external sources (CSV, QTI, Blackboard export, Quizizz export) and export Conker quizzes to standard formats for portability. The system parses source formats, maps question types and metadata to Conker's internal schema, and handles format conversion (e.g., Blackboard XML → Conker JSON). Likely uses format-specific parsers and a canonical internal representation for flexible import/export.

Enables multiple educators to work on the same quiz simultaneously, with real-time collaboration features (live editing, presence indicators), version history tracking, and comment threads on specific questions. The system likely uses operational transformation (OT) or conflict-free replicated data types (CRDTs) for concurrent editing, stores version history in a database, and provides rollback capability to previous versions.