PlotProof for MATLAB

Note: This analysis is based entirely on the self-reported project description provided. No external verification, traction data, revenue figures, customer names, or independent sources are available. All claims are treated as stated by the author.

The description states that PlotProof for MATLAB:

• Reads CSV, Excel, or MAT data
• Uses a MATLAB plotting engine to render two to five ranked chart candidates
• Applies GPT-5.6 Sol in Codex to review candidate images against research goals
• Scores claims on support, legibility, accessibility, honesty, and reproducibility
• Enforces strict schema validation before MATLAB starts
• Allows only five controlled figure repairs (no executable MATLAB code)
• Exports PNG/SVG figures, before-and-after comparisons, Markdown/JSON evidence, and a self-contained HTML decision report
• Includes a manifest recording byte count and SHA-256 digest of source artifacts for tamper-proofing

Inference: The product is not a general-purpose charting tool but a specialized workflow for scientific visualization review with AI-assisted quality control.

The author claims:

• A scientific chart can be polished and still be misleading
• Rule-based recommenders are useful but assign meaning the data does not contain
• The goal is to go beyond generating figures — challenge choices, leave evidence for every decision

Claim: This positions PlotProof as a tool for improving reproducibility and transparency in scientific visualization workflows.

Inference: It evolved from a hackathon idea into a structured system with validation layers, adversarial testing, and offline reporting capabilities.

The description does not explicitly name target customers or personas. However:

• The tool is built for MATLAB users who generate scientific charts
• It targets researchers or data scientists working in environments where chart accuracy and reproducibility matter
• The focus on controlled repairs and evidence suggests it may appeal to those needing audit trails or compliance

Not evidenced: No explicit customer segments, use cases, or ICP defined.

There is no mention of pricing, monetization strategy, or business model in the description.

Not evidenced: No indication of how this would be sold or whether it has a commercial model.

The author states:

• Built with bash, codex, github, json, matlab, openai, python
• Uses MATLAB R2025a
• Implements Codex as primary development environment
• Includes contract tests, validation gates, adversarial benchmarking
• Has a release gate including shell, unit, visual, privacy, provenance, and GitHub Actions checks
• Demonstrates offline HTML report generation with embedded evidence

Inference: The technical stack shows integration of AI (Codex/GPT), MATLAB, and CI/CD practices. It suggests a focus on robustness and reproducibility.

The description includes:

• One-command demo workflow
• 15/15 reproducible contract checks
• All 14 adversarial outputs fail closed
• SHA-256 integrity verification across 12 real MATLAB evidence artifacts
• A GitHub repository with documented development history and test results

Not evidenced: No user feedback, customer base, or adoption metrics beyond the author’s own testing.

The description does not reference competitors or existing tools in the space. It implies a niche for AI-assisted scientific chart review, especially where reproducibility and auditability are important.

Not evidenced: No competitive landscape or market positioning provided.

• The project is self-reported by one person (William Isode) with no third-party validation
• No evidence of real-world usage or customer feedback
• No mention of scalability, integration with other tools, or enterprise readiness
• The tool is tied to a specific MATLAB version (R2025a), which may limit adoption
• The claim that GPT-5.6 Sol reviews images and scores them raises questions about image-to-text accuracy and consistency

Inference: The project appears experimental and limited in scope, with no clear path to commercial viability or widespread use.

1. What is the actual workflow for integrating this into a research lab or engineering team?
2. How does it handle edge cases not covered in the demo?
3. Are there plans to support other plotting environments beyond MATLAB?
4. Has anyone outside of the author tested or used this tool?
5. What are the long-term maintenance and upgrade plans for MATLAB compatibility?