Artifact Rescue

The description states that Artifact Rescue is a tool that inspects Excel workbooks (.xlsx), extracts formulas, comments, and behavior, and generates a Recovery Compatibility Report and Rule Ledger. It uses GPT-5.6 to clarify ambiguous rules but does not treat AI output as the final authority.

It builds a modern application from recovered rules, including fields, role-gated data entry, calculations, reports, and CSV export. The process includes:

• A Proof Room that tests legacy and restored behavior side-by-side
• Transactional migration into Supabase tables with checksum reconciliation and rollback support
• A final package including the rebuilt app, migrated records, operating guide, and certificates

The tool is described as a restoration studio for dying software, focused on institutional knowledge preservation.

Evidence Self-reported. No external validation or product screenshots provided.

The author positions Artifact Rescue as a solution to the problem of institutional knowledge loss when employees leave organizations that depend on spreadsheets or desktop tools.

Key claims:

• It restores abandoned spreadsheets into modern applications.
• It proves every recovered rule before migration.
• It avoids “convincing but untrustworthy” demos by using deterministic checks and human review.
• AI is used for clarification, not as the trust boundary.
• The core value proposition is proof, not generation.

The positioning evolved from a hackathon prototype to an idea of a restoration studio that can be applied across domains (e.g., food banks, animal shelters, clinics).

Evidence Self-reported. No prior version or market positioning history provided.

The description states that the tool is intended for small organizations that depend on spreadsheets or desktop tools built by one person or volunteer. These are typically non-profits or small teams where knowledge is not codified and can be lost when people leave.

It also implies a use case in domains with legacy workflows, such as:

• Food banks
• Animal shelters
• Clinic scheduling

The ICP appears to be:

• Small, non-technical teams
• Organizations with institutional knowledge embedded in spreadsheets
• Users who want to migrate data and behavior without losing trust or accuracy

Evidence Self-reported. No customer data, personas, or segmentation provided.

No business model or pricing information is provided in the description.

The author states that the tool is a prototype, built for a hackathon, with no mention of monetization, subscriptions, or licensing.

Evidence Not evidenced.

The product is described as:

• Built using Next.js and deployed on Vercel
• Backed by Supabase Postgres
• Uses ExcelJS to parse workbooks
• Employs GPT-5.6 via OpenRouter for clarification, with Zod for validation
• Uses Playwright for browser testing and GitHub Actions for CI/CD
• Includes a generic formula interpreter that supports arithmetic, comparisons, IF logic, aggregates, and constrained VLOOKUP/XLOOKUP
• Has a Proof Room, regression suite, and rollback support

The system is described as:

• Deterministic where possible
• Reversible
• Schema-validated
• Designed to avoid “inventing” uncertainty or silent guessing

Evidence Self-reported. No performance data, scalability claims, or delivery metrics.

No traction or maturity signals are evident in the description:

• No customers
• No revenue
• No product usage metrics
• No prior versions or iterations
• No public deployment beyond a hackathon submission

The project is described as a single-person hackathon effort, with no indication of ongoing development, user feedback, or adoption.

Evidence Not evidenced.

No competitive landscape is described. The author does not reference existing tools for spreadsheet migration or knowledge preservation.

The product appears to be unique in its focus on:

• Proving recovered rules
• Using AI as a clarification tool, not a trust boundary
• Supporting transactional migration and rollback

Evidence Not evidenced.

1. Unproven demand: The project is described as a hackathon prototype with no evidence of real-world adoption or customer need.
2. Single-founder build: Only one person built the product, which raises questions about scalability and long-term maintenance.
3. AI dependency without clear trust model: While GPT-5.6 is used for clarification, it is not the trust boundary — but this may still be a point of risk if AI output is misused or misunderstood.
4. No monetization strategy: No indication of how the product would generate revenue.
5. Limited scope: The author notes that the tool supports a “useful subset” of Excel functions and does not cover all edge cases.

Evidence Inferred from self-reported description.

1. What is the real-world problem you are solving, and how did you identify it?
2. Have you tested this with any actual users or organizations?
3. How do you plan to scale beyond a single-person prototype?
4. What is your path to monetization?
5. How do you intend to expand beyond the current Excel function support?
6. What are the key assumptions in your approach, and how might they fail?