UnAI Evidence — Explainable Portrait Analysis

The description states that UnAI Evidence is a tool for analyzing portrait images and presenting visible editing-related signals in an explainable way. It works by:

• Accepting user-uploaded portraits with explicit consent;
• Locally validating the image and creating a cost-limited 1024 px analysis copy;
• Measuring lightweight deterministic signals such as edge density, luminance spread, saturation, and compression hints;
• Sending this data to GPT-5.6 Sol for structured report generation;
• Presenting results visually with thin arrows pointing to evidence locations and expandable explanations;
• Optionally offering a separately consented GPT Image 2 edit that creates one clearly labeled naturalized interpretation.

The tool is described as deliberately not claiming to recover the true original, nor to prove that an image is fake. It is an explainability aid for visible signals only.

The author states that the tool asks a more useful question than traditional detection tools: “what visible editing-related signals are present, where are they, and what ordinary camera or processing effects could explain them?”

It positions itself as an alternative to systems that reduce uncertainty to one unexplained percentage, aiming instead for transparency in its findings.

The description does not explicitly name target customers or define an ideal customer profile (ICP). However, it implies usage by individuals who upload portraits and seek clarity on editing signals.

There is no evidence of pricing, monetization strategy, or business model in the description. The tool appears to be a prototype built for a hackathon and lacks any indication of commercial use.

The tool is built using:

• TypeScript/React with TanStack Start and Tailwind CSS;
• Server functions via OpenAI Responses API;
• GPT-5.6 Sol for structured reasoning;
• GPT Image 2 for optional reinterpretation;
• Codex as the primary development environment.

It uses deterministic measurements and a strict JSON schema to guide GPT outputs, separating evidence from confidence and ensuring alternative explanations are included.

There is no evidence of traction or adoption beyond the author’s own development. No customer data, usage metrics, or user feedback are provided.

The description does not provide information about competitors or the competitive landscape. It only contrasts UnAI Evidence with existing tools that “reduce uncertainty to one unexplained percentage.”

1. No commercial traction or adoption: The tool appears to be a prototype without any evidence of real-world use.
2. Unverified claims about accuracy: The author states that it cannot reconstruct edit history or prove an image is fake, but there is no validation of its output quality or reliability.
3. Limited scalability: The architecture is cost-limited and uses only one GPT-5.6 analysis per report.
4. Dependency on proprietary APIs: Heavy reliance on OpenAI models (GPT-5.6 Sol, GPT Image 2) may pose risks if those services change or become unavailable.

1. What specific use cases are you targeting beyond the hackathon prototype?
2. Have you tested the accuracy of GPT-5.6 Sol’s structured outputs in real-world scenarios?
3. How do you plan to scale or monetize this product if it gains traction?
4. Are there any known limitations or biases in how the deterministic measurements are interpreted?
5. What is your roadmap for integrating multiple specialist models and calibrating them?
6. Have you considered legal or ethical implications of labeling images as “edited” or “synthetic”?