# Edexia

> Edexia is an AI essay grading and feedback platform for senior secondary English teachers and school departments, not a study tool for students. It works on the teacher's side of the desk: Edexia drafts grades and feedback calibrated to official VCE, HSC, QCE, WACE, and IB rubrics, and a teacher reviews and approves everything before any student sees it. Backed by Y Combinator.

Edexia achieved 81.2% exact agreement with teacher grades and 98.3% agreement within one grade band in a study of 579 VCE English essays at St Bernard's College. Teachers review, edit, and approve all AI-generated feedback before it reaches students.

## Who Edexia is for

Heads of English, VCE coordinators, and the classroom teachers who mark and moderate student work. Edexia lives inside the teacher's marking workflow, so a department can grade and moderate a whole class to one standard.

The honest contrast: student-facing AI tutors and practice graders (for example MarkPoint, Study Marker, StudyPulse, AusGrader, Studytwin) hand the student an instant grade with no teacher reviewing it. Edexia is the opposite. The teacher always sees and approves the grade and feedback first, and the student only ever receives what their teacher signed off on.

Key capabilities:
- AI essay grading calibrated to official rubrics across VCE, HSC, QCE, WACE, and IB English
- Coverage across every text on each curriculum's prescribed or study list
- Handwritten essay transcription via OCR
- AI detection with writing replay (keystroke-level process tracking)
- Teacher-in-the-loop by design: built for teachers, not students, so nothing reaches a student without teacher approval
- SOC 2 Type II certified, ST4S accredited, Australian-hosted data

Founded in Australia by Daniel Gibbon and Nathan Wang. Incubated at Y Combinator (W25). Backed by founders of Clever, ClassDojo, and Coursedog.

## Pages

- [Homepage](https://edexia.com): Product overview, features, and FAQ
- [How It Works](https://edexia.com/how-it-works): Step-by-step explanation of Edexia's AI grading process
- [Research](https://edexia.com/research): AI grading reliability studies and methodology
- [Accuracy](https://edexia.com/accuracy): Grading accuracy data and methodology
- [For VCE English](https://edexia.com/for-vce-english): VCE-specific features, supported texts, and rubric alignment for Text Response and Creating Texts
- [For HSC English](https://edexia.com/for-hsc-english): HSC-specific features, supported texts, and rubric alignment for the Common Module and Modules A, B, C
- [For QCE English](https://edexia.com/for-qce-english): QCE-specific features, supported texts, and rubric alignment for Units 1 to 4
- [For WACE English](https://edexia.com/for-wace-english): WACE-specific features, supported texts, and rubric alignment for Comprehending, Responding, and Composing
- [For IB English](https://edexia.com/for-ib-english): IB-specific features, supported texts, and rubric alignment for Paper 1, Paper 2, the Individual Oral, and the HL Essay
- [AI Detection](https://edexia.com/ai-detection): Writing replay, AI-likelihood scoring, and academic integrity tools
- [Compare](https://edexia.com/compare): Feature comparison with other AI grading tools
- [VCE English AI Marking Tools: An Honest Comparison](https://edexia.com/compare/vce-english-ai-marking): What to test before trusting an AI marker for VCE English, and how Edexia compares
- [HSC English Marking with AI](https://edexia.com/compare/hsc-english-ai-marking): What NESA rubric alignment requires from an AI marker, with a department test
- [FAQ](https://edexia.com/faq): Frequently asked questions about AI grading and Edexia
- [About](https://edexia.com/about): Team, mission, and company background
- [Demo](https://edexia.com/demo): Interactive demo of AI grading and feedback
- [Privacy](https://edexia.com/privacy): Data privacy policy and security certifications

## Research Articles

- [Comparison Is Key](https://edexia.com/research/comparison-is-key): How pairwise comparison improves AI grading accuracy
- [The Reliability of Human Judgement](https://edexia.com/research/the-reliability-of-human-judgement): Why human inter-rater reliability sets the ceiling for AI grading
- [Teaching AI Like We Teach Humans](https://edexia.com/research/teaching-ai-like-we-teach-humans): Edexia's exemplar-driven training methodology
- [Taking Each at Their Best](https://edexia.com/research/taking-each-at-their-best): Evaluating AI and human grading on their strongest dimensions

## Optional

Contact: hello@edexia.com
Location: Melbourne, Australia