CarryForward

Note: This analysis is based entirely on the self-reported project description provided by the authors. No third-party verification, historical data, or independent sources are available.

The description states that CarryForward:

• Is an AI-powered tool for checking document readiness before submission.
• Allows users to specify what requirements are needed (e.g., name, date of birth).
• Accepts PDF, PNG, or JPEG uploads.
• Uses Gemini 2.5 Flash to extract visible facts from documents.
• Applies rule-based logic in TypeScript to determine if the extracted data matches, conflicts, is missing, or is unclear relative to user-defined requirements.
• Provides a Plan B and editable email template for next steps.

It also states that:

• It does not verify document authenticity.
• It does not make eligibility or admission decisions.
• It separates AI extraction from comparison logic to ensure consistency and clarity for users.

Inference: The product is described as a web-based readiness-checking tool, not an identity verification or decision-making platform. Its core function appears to be fact-matching across documents using AI + deterministic rules.

The authors state:

• CarryForward was inspired by practical problems in rural areas of Pakistan where small paperwork errors cause significant delays and stress.
• It aims to help users find issues before submission, rather than after rejection or delay.
• The tool is positioned as a readiness-checker for applications like university admissions or scholarships.

Inference: The positioning evolved from addressing a local problem (paperwork mismatches in developing countries) to a broader use case of document validation prior to application submission. However, the description does not indicate any shift toward targeting global markets or enterprise clients.

The description states:

• The tool is intended for individuals applying to universities, scholarships, or other formal processes.
• It targets people in developing countries (specifically Pakistan) who face document-related issues due to spelling mismatches or incomplete paperwork.
• Users define what the application requires and upload their own documents.

Inference: The ICP appears to be individual applicants with limited access to support systems, particularly those dealing with bureaucratic processes that require strict documentation. No evidence of institutional or enterprise adoption is presented.

The description does not contain any information about:

• Revenue streams
• Pricing models
• Monetization strategy
• Subscription plans or usage fees

Not evidenced: There is no indication of how the tool would generate revenue or be monetized in a commercial context.

The authors state:

• Built with React, TypeScript, Vite, and Cloudflare Workers.
• Uses Gemini 2.5 Flash for document extraction.
• Employs Codex and GPT-5.6 during development (not used at runtime).
• Implements private R2 storage for temporary files.
• Uses Zod for data validation.
• Includes error handling for file types, network failures, and invalid content.

Inference: The technical stack suggests a modern web application with backend API routes, AI integration, and basic security practices. However, no evidence of production-grade infrastructure or scalability is provided.

The description states:

• The project was built by second-semester students.
• It is described as a working prototype that could be integrated into real systems.
• No mention of users, customers, revenue, or product usage metrics.
• The team tested the tool with fictional records and demo videos.

Not evidenced: There is no evidence of traction, user base, or market validation beyond internal testing and a hackathon submission.

The description does not include:

• Information about competitors
• Market analysis
• Prior art or similar tools in the space

Not evidenced: No competitive landscape or differentiation from existing solutions is described.

Key risks identified from the self-reported description:

1. Lack of commercial traction or adoption — no evidence of users, customers, or revenue.
2. Limited scope and language support — currently only supports English documents.
3. Dependency on AI model selection and cost awareness — the team mentions challenges with model choice and token usage.
4. No clear path to monetization or scaling — no indication of how the tool would be commercialized.
5. Prototype-level maturity — built by students, not a mature product or company.

Inference: The project is at an early stage and lacks evidence of viability as a scalable business model or product.

1. What specific document types or application domains are you targeting beyond English documents?
2. How do you plan to handle multilingual support and regional variations in document formats?
3. Have you tested the tool with real users outside of your own team or hackathon environment?
4. Is there a roadmap for integrating with universities, scholarship platforms, or online portals?
5. What are the key assumptions about user behavior and adoption that underpin your product design?
6. How do you intend to manage costs associated with AI APIs and backend infrastructure at scale?
7. Are there any legal or compliance considerations around handling personal documents in your system?