CoTeacher

Company: CoTeacher

Self-reported purpose: AI-assisted grading for K–12 teachers — scan assessments, verify answers, turn results into actionable student feedback, and draft parent updates while teachers retain final control.

Key claim: AI assists. Teacher decides.

What changed: The project is a hackathon prototype built to demonstrate an offline, browser-based OMR system with a five-step workflow for assessment grading and feedback generation. It does not include real AI inference at runtime, nor does it connect to any student database or communication platform.

Single most important open question: What is the actual product-market fit of this solution in real K–12 classrooms? The description states no revenue, customers, or adoption data — only a demo.

The description states that CoTeacher is a responsive React 19 + Vite + TypeScript application deployed on Vercel. It implements a five-step workflow for teachers to:

1. Set up an exam with answer keys and learning targets.
2. Scan or upload assessments using an OMR sheet.
3. Review ambiguous or short-answer responses manually.
4. Generate teaching insights based on verified responses.
5. Draft parent updates, which require explicit approval.

The system uses browser-local OMR processing, including registration-marker detection, planar homography normalization, and bubble classification — all done without uploading images or using external AI models at runtime. Short answers are flagged for manual review; no handwriting recognition is claimed.

Inference: The product is a demo-grade prototype built to showcase how an AI-assisted grading workflow could function in K–12 settings. It does not yet include real integrations with student information systems (SIS), AI inference, or communication platforms.

The description states that CoTeacher began with the principle: AI assists. Teacher decides. This is a stated positioning and trust model — not a fact about traction or adoption.

It also says the project was built to reduce repetitive administration for teachers, focusing on grading, pattern identification, and follow-up instruction. The product is positioned as a tool that preserves teacher judgment, rather than replacing it.

Inference: The positioning reflects an intent to build a teacher-first AI assistant, not a fully autonomous grading system. It emphasizes control and transparency over AI use.

The description states that CoTeacher is for K–12 teachers, specifically those who grade paper assessments and want to reduce time spent on repetitive tasks like marking, identifying learning patterns, and writing family updates.

Inference: The target customer is a teacher in a K–12 environment, likely in a classroom setting with paper-based assessments. The ICP (Ideal Customer Profile) appears to be teachers who are early adopters of AI tools and value control over grading decisions.

The description does not state any business model or pricing strategy. It says the demo is fully offline, with no external calls to AI, analytics, or messaging services. The project was built for a hackathon, not as a commercial product.

Inference: There is no evidence of a business model or pricing structure. The prototype does not indicate any monetization strategy or customer acquisition plan.

The system is built with:

• React 19 + Vite + TypeScript
• Responsive web design
• Browser-local OMR processing using Canvas and ImageData
• Pure TypeScript implementation for OMR core
• Context/reducer pattern for workflow state management
• No external AI or data storage at runtime

The description says that the OMR path detects registration markers, normalizes images with homography, and classifies bubbles locally, without uploading or storing image data.

Inference: The technical stack is modern and lightweight, designed for browser-based execution. It shows a strong focus on privacy and local processing, but lacks real-world integration or scalability signals.

The description states that this is a hackathon demo. It does not include any evidence of:

• Revenue
• Customers
• Adoption
• Product-market fit
• Real usage data
• Production deployment

Inference: There are no traction or maturity signals beyond the prototype itself.

The description does not mention any competitors or market context. It is a self-contained narrative focused on the product’s design and functionality, without reference to existing tools in the K–12 grading or AI-assisted education space.

Inference: No competitive landscape is described. The project appears to be independent of known players, but this cannot be confirmed due to lack of evidence.

• No real-world usage or feedback: This is a demo, not a product in use.
• No AI inference at runtime: The system does not use external AI models for scoring or insights — it only simulates them.
• No integration with SIS or communication platforms: No real data flow or user-facing integrations.
• No pricing or monetization model: No indication of how the product would be sold or used commercially.
• No evidence of teacher adoption or feedback loops: The project does not describe any user testing or real-world validation.

Inference: The biggest risk is that this is a conceptual prototype, not a viable product. It may not have sufficient commercial traction or market demand to justify further investment.

1. What specific K–12 classroom use cases were you targeting, and how did you validate them?
2. How would the system integrate with existing SIS or LMS platforms?
3. What are your plans for incorporating real AI inference into scoring or feedback generation?
4. Have you conducted any user testing with teachers in actual classrooms?
5. What is the path to production? What features are planned beyond this demo?