// AI.STUDIO

    AI Studio — Content Generation Infrastructure for Education

    Worksheet Builder and Lesson Plan Builder — two API suites that generate calibrated practice content and instructional plans from the same evaluation signals.

    Explore APIs →
    // AI.STUDIO.SESSION
    active_suites:[WORKSHEET.BUILDER, LESSON.PLAN.BUILDER]
    gap_signal_source:EVALUATION.INFRASTRUCTURE ✓
    content_generated:2,847 this session
    review_queue:12 PENDING
    HUMAN.REVIEW:REQUIRED
    signal_accuracy:93% ICC
    SYSTEM.STORES:LEARNING.SIGNALS ✓
    // GOVERNANCE.BY.DESIGN
    5-STEP WIZARD
    Teacher Configures Before Generation Runs
    →Permitted formats + difficulty ranges set
    →Topic constraints + gap severity thresholds
    →No generation outside your defined policy
    PENDING_HUMAN
    API-Enforced Review on Every Generation
    →review_status: PENDING_HUMAN on every request
    →Cannot deliver until reviewer approves
    →No bypass parameter exists
    FULL AUDIT TRAIL
    Signal-to-Content Traceability
    →Triggering evaluation signal + gap topic
    →Reviewer ID + approval/rejection decision
    →Generation + delivery timestamps logged
    // THE.SYSTEM

    Two Suites, One Source of Truth

    AI Studio sits downstream of evaluation — after signals are captured, gaps are identified, and structured learning intelligence is produced.

    // WORKSHEET.BUILDER
    endpoints:2
    generate-questions:gap_topic → practice questions
    generate-feedback:response → structured feedback
    signal_source:EVALUATION.INFRASTRUCTURE ✓
    review_gate:PENDING_HUMAN ✓
    // LESSON.PLAN.BUILDER
    endpoints:1
    generate-lesson-plan:gaps → instructional plan
    sequencing:GAP.PRIORITY (not syllabus order)
    signal_source:EVALUATION.INFRASTRUCTURE ✓
    review_gate:PENDING_HUMAN ✓
    // both.suites / same.signal.source / human.review.required
    // CAPABILITIES

    Worksheet Builder and Lesson Plan Builder

    WORKSHEET.BUILDER
    Gap-Targeted Question Generation
    →Calibrated to gap topic + difficulty level
    →Tied to evaluation signal, not topic category
    →POST /worksheetbuilder/v1/generate-questions
    WORKSHEET.BUILDER
    Structured Feedback Generation
    →Aligned to evaluation criteria + gap addressed
    →90%+ accuracy, delivered instantly
    →POST /worksheetbuilder/v1/generate-feedback
    LESSON.PLAN.BUILDER
    Gap-Targeted Lesson Plans
    →Sequenced by gap priority, not syllabus order
    →Full pedagogical structure for teacher delivery
    →POST /lessonplanner/v1/generate-lesson-plan
    SIGNAL.TRIGGER
    No Manual Brief Required
    →Gap topic + severity from memory layer
    →Learner context flows from evaluation signals
    →No manual content brief required
    REVIEW.GATE
    API-Enforced Human Approval
    →PENDING_HUMAN status on every request
    →Cannot deliver until reviewer approves
    →No bypass parameter exists
    JSON.OUTPUT
    Structured Metadata on Every Response
    →Concept tags + difficulty + gap alignment
    →Review status + reviewer ID included
    →Ready for platform content delivery
    // HOW.IT.WORKS

    From Evaluation Signal to Generated Content

    01
    Receive Gap Signal
    Gap topic, severity, and learner context from evaluation
    02
    Select Suite and Generate
    Worksheet or Lesson Plan API call with gap inputs
    03
    Human Review
    PENDING_HUMAN queue — teacher approves, edits, or rejects
    04
    Structured Delivery
    JSON output with gap alignment, concept tags, reviewer audit
    // POST /worksheetbuilder/v1/generate-questions
    // Request
    {
    "learner_id": "std_00391",
    "gap_topic": "organic_chemistry",
    "gap_severity": "high",
    "question_count": 5,
    "difficulty": "moderate",
    "format": "short_answer"
    }
    // Response
    {
    "generation_id": "gen_84291",
    "status": "PENDING_HUMAN",
    "questions": [
    {
    "question_id": "q_001",
    "text": "Describe the mechanism of an SN2 reaction...",
    "concept_tag": "nucleophilic_substitution",
    "gap_alignment": "organic_chemistry_mechanism",
    "difficulty": 3.1
    }
    ],
    "review_required": true
    }
    // THE.SYSTEM

    Where AI Studio Fits in the Infrastructure

    UPSTREAM

    Evaluation System

    • →Signal capture + scoring engine
    • →Generates gap data AI Studio consumes
    • →Infrastructure layer AI Studio draws from
    See the product →
    API LAYER

    Evaluation APIs

    • →REST APIs for signal capture and scoring
    • →gap_topic + gap_severity as structured outputs
    • →Upstream input layer AI Studio reads from
    See the APIs →
    CONTENT LAYER

    Content Generation

    • →Worksheet Builder + Lesson Plan Builder
    • →Signal-triggered, human review gated
    • →The content pipeline AI Studio powers
    See content generation →
    // PRODUCTION.PROOF

    Content Built on Validated Evaluation Signals

    93%
    Scoring Reliability (ICC)
    →Evaluation signals driving AI Studio are 93% reliable
    →Signal quality determines content calibration quality
    →Evaluation infrastructure built before content generation
    90%+
    Feedback Accuracy, Delivered Instantly
    →Same evaluation engine powers feedback generation
    →Same system, same accuracy, same signal source
    →Structured, criteria-aligned, not generic
    Up to 60%
    Reduction in Evaluation Time
    →Phase 1 validated — India AI Impact Summit 2026, MeitY
    →Time recovered becomes capacity for content review
    →AI Studio makes recovered time productive
    // GOVERNANCE

    Human-Controlled Generation at Every Layer

    No content generated by AI Studio reaches a learner without human approval — this is a structural control point enforced at the API level. Academic teams configure content policies; teachers own what learners see.

    Human Review Gate (API-Enforced)
    →review_status: PENDING_HUMAN on every POST request
    →Cannot deliver until reviewer explicitly approves
    →No bypass parameter — structural, not optional
    Content Policy Configuration
    →Permitted formats + difficulty bands
    →Topic constraints + gap severity thresholds
    →System generates strictly within configured policy
    Full Audit Trail
    →Triggering evaluation signal + gap topic
    →Reviewer ID + approval/rejection decision
    →Generation + delivery timestamps logged
    // INTEGRATION.SURFACE

    Two API Suites, Production-Ready

    →Both suites return structured JSON
    →Both require PENDING_HUMAN review before delivery
    →Standard inputs: gap_topic · gap_severity · learner_id or cohort_id
    Explore APIs →
    # Worksheet Builder — 2 endpoints
    POST /worksheetbuilder/v1/generate-questions # Practice questions
    POST /worksheetbuilder/v1/generate-feedback # Structured feedback
    # Lesson Plan Builder — 1 endpoint
    POST /lessonplanner/v1/generate-lesson-plan # Instructional plan
    # Standard inputs (all endpoints):
    # gap_topic · gap_severity · learner_id (or cohort_id)
    # All responses include:
    # review_status: PENDING_HUMAN

    // COMMON.QUESTIONS

    CGF AI Studio is an education content generation product — two API suites that generate gap-targeted practice questions, feedback, and lesson plans from evaluation signals. Every piece is generated within academic policies configured by teachers, then held in human review before delivery. Evaluation signals feed the generation. Human approval gates the output.

    What is CGF AI Studio?

    AI Studio is CGF's content generation product — the system that generates practice questions, structured feedback, and lesson plans from evaluation signals. It contains two API suites: Worksheet Builder (gap-targeted questions and feedback for learners) and Lesson Plan Builder (structured instructional plans for teachers). All generated content requires human review before delivery.

    How does AI Studio connect to the evaluation infrastructure?

    AI Studio sits downstream of CGF's evaluation infrastructure. The gaps identified during evaluation — captured as structured learning intelligence — become the inputs for content generation. Gap topic, severity, and learner context flow from the same memory layer that evaluation builds over time. Content generation without evaluation signals is generic. CGF's is not.

    Is AI Studio autonomous or human-controlled?

    Human-controlled at every layer. The review gate is enforced at the API level — generated content carries review_status: PENDING_HUMAN and cannot be delivered until a designated reviewer approves it. Academic teams configure the content policies. Teachers own what reaches learners. AI Studio is infrastructure for human decision-making, not a replacement for it.

    What is the accuracy of the evaluation signals that drive AI Studio?

    The evaluation infrastructure that feeds AI Studio achieves 93% scoring reliability (ICC) and delivers 90%+ accurate feedback instantly. Content generation is downstream of evaluation — if the signal is unreliable, the content is miscalibrated. CGF builds the evaluation infrastructure to production accuracy before content generation begins.

    How do I integrate AI Studio into my platform?

    AI Studio exposes two REST API suites — Worksheet Builder and Lesson Plan Builder. Both accept standard inputs (gap_topic, difficulty, learner_id) and return structured JSON with metadata including concept tags, gap alignment, and review status. Full integration documentation is available at docs.crazygoldfish.com. For implementation support, book a demo.

    // NEXT.STEP

    Build Content Generation That Knows What Learners Need

    Gap-driven questions, feedback, and lesson plans — generated from evaluation signals, reviewed by humans before delivery.

    Explore APIs →
    See how evaluation becomes infrastructure.