ColdOpen - Teaching Computer Skills Made Easy

The description states that ColdOpen is a Chrome extension plus a teacher page. The extension observes interaction timing and event shape on classroom Chromebooks and laptops, such as hesitation before clicks, abandoned drags, file dialogs opened and dismissed without selection, and keystroke intervals.

It does not capture content — no screenshots, keystrokes, filenames, or page text. This is described as a structural property of the code, with tests in the repository that fail the build if content scripts reference DOM text, selectors, or key identity.

The extension aggregates individual session data into a class-level feature vector. Identity is discarded at aggregation; there are no per-student records.

The teacher page uses the OpenAI API to turn the class vector into a five-minute demo script — including diagnosis, setup, steps, and a comprehension check. The model call happens server-side in a Cloudflare Pages Function.

The author states that ColdOpen is positioned as an alternative to punitive monitoring software used by districts. It is described as a diagnostic tool, not a surveillance tool — aiming to help teachers understand what students are missing rather than catch them doing something wrong.

The system is built around the idea that teachers cannot act on individual student struggles due to time constraints, so it aggregates data at class level. This is framed as both a privacy win and a product necessity.

The author also notes that ColdOpen was designed with equity in mind, recognizing that many classrooms use older hardware that cannot support on-device AI. The solution is architectural: minimal device-side processing, server-side model inference.

The description states that the target customer is teachers in Colorado, particularly those working with students who lack exposure to computer skills due to socioeconomic factors. These are described as students who arrive at school without basic computer fluency, often because they do not have access to computers at home.

The system is designed for classroom environments using Chromebooks or laptops, where teachers have limited time and resources to intervene on individual students.

There is no mention of specific school districts, grade levels, or types of schools beyond the context of Colorado and under-resourced classrooms.

Not evidenced. The description does not contain any information about pricing, monetization, or business model.

The system is built using:

• MV3 Chrome extension
• TypeScript for aggregation and generation
• Zod contracts shared across the pipeline
• Cloudflare Pages plus one Pages Function
• Static fixtures committed to the repo
• A test suite covering contracts, rollup, privacy invariant, and deployed function

The author emphasizes that the system was built spec first, agent second, with a design document specifying non-goals such as:

• No per-student state
• No content capture
• Key server-side only

A key constraint — that the extension cannot read content — is enforced by a test in the repository that checks the source code for references to DOM text or key identity.

Not evidenced. The description does not include any data on:

• Revenue
• Customers
• Usage metrics
• Adoption rates
• Product maturity beyond prototype status

The author notes that real classroom traces were not collected due to timing (school ends in July), and the demo runs on fixtures they authored.

The description states that district surveillance software exists for the same need but is punitive and despised by teachers. ColdOpen is positioned as an alternative that flips the gaze from catching to diagnosing.

It also notes that Google’s built-in AI requires Chromebook Plus hardware, which excludes many classrooms. This creates a structural gap in the market that ColdOpen aims to fill with its architecture.

No other competitors are named or described.

• Empirical validation is missing: The system's effectiveness depends on real student behavior separating into distinct patterns, but this has not been demonstrated beyond fixtures.
• No product-market fit data: There is no evidence of traction, customer feedback, or usage metrics.
• Single-founder project: The team size is listed as one (Kevin Loeffler), which may limit execution capacity.
• Privacy architecture is self-enforced: While described as robust, it relies on code-level enforcement rather than external auditing or compliance frameworks.
• Unclear scalability assumptions: The system aggregates at class level, but how this scales to larger groups or districts is not discussed.

1. What empirical evidence do you have that real students' behavior separates into the patterns your system expects?
2. How do you plan to validate the effectiveness of demo scripts generated by the AI in actual classroom settings?
3. Are there any known technical or legal barriers to deploying this at scale in schools or districts?
4. What are the implications of relying on a single developer for both product design and implementation?
5. Have you considered how teachers might integrate ColdOpen into their existing workflows, especially given limited time and resources?
6. How do you intend to address potential resistance from school administrators who may not see value in this type of tool?