Fabric: Verified AI-Composed Interfaces

This is a self-reported project with no evidence of revenue, customers, traction or independent verification. The author claims to have built a system that separates creative presentation from deterministic guarantees, but the description does not substantiate whether this separation is technically feasible or practically effective.

The description states Fabric is:

• An AI-native Service-to-UI system for verified, AI-composed interfaces
• A system where services publish structured contracts describing capabilities, data semantics, required disclosures, confirmation rules, permissions and constraints, valid state transitions, and cross-surface obligations
• A system that uses GPT-5.6 to compose adaptive interfaces across ChatGPT, web, and voice surfaces
• A system that runs deterministic conformance checks against structured evidence bindings to verify required information survives
• A system with a shared state engine preventing agents from bypassing required steps like explicit confirmation

The author describes Fabric as not being:

• A JSON component renderer (which describes widgets like buttons and forms)
• Merely generative UI (which allows AI to adapt layout and wording but not remove provider requirements)

The description states the author's positioning:

• AI is moving software beyond fixed pages and predefined workflows
• Users will increasingly interact through conversation while visual and voice interfaces are composed dynamically around the task
• Fabric explores a third model: service defines what must survive, AI decides how it appears, and Fabric proves the result

The claim evolution shows:

• From traditional approaches (hand-built widgets with control but duplication across surfaces)
• To current approach (AI composes adaptive interfaces but risks losing critical disclosures)
• To Fabric's approach (structured contract defines requirements, AI composes presentation, deterministic checks ensure survival)

The author claims Fabric is not just a rendering format but a system that distinguishes what AI may creatively adapt from what the provider requires and what runtime can verify.

Not evidenced. The description does not state who the target customers are or what their specific needs are beyond general service providers who want to compose interfaces across multiple modalities while maintaining compliance requirements.

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

The description states Fabric's technical components include:

• Fabric Contract schema
• Typed Experience IR (Intermediate Representation)
• GPT-5.6 structured composition
• Deterministic conformance rules
• Evidence bindings
• Generation and repair loop
• Shared state-transition engine
• ChatGPT Apps SDK target
• Web renderer
• Voice or conversational renderer
• Live conformance and semantic event panels

The author states:

• The prototype is being developed as a small vertical slice rather than complete universal interface standard
• Core components are described in the architecture section
• Challenges include separating creative presentation from deterministic guarantees
• Focus for prototype is on machine-checkable guarantees like required semantic content, evidence binding, modality coverage, disclosures before specified states, confirmation events, invalid transitions, and verbatim legal content

Not evidenced. The description contains no information about:

• Revenue or customers
• Product usage metrics
• Market traction
• Adoption rates
• Customer feedback
• Product maturity beyond prototype status

The author only describes a prototype demo scenario with deliberate attempts to omit required information, which was rejected by the system.

Not evidenced. The description does not mention any competitors, market positioning relative to existing solutions, or competitive landscape analysis.

1. What is the actual technical implementation of the conformance checking? How does it determine what constitutes "required information" vs. "creative adaptation"?

1. Can you demonstrate a working example where the system successfully handles a complex interface requirement that would be difficult for AI to get right without deterministic checks?

1. How do you plan to scale this from a prototype built by one person to a commercial solution?

1. What specific problems are you solving that existing UI frameworks or interface composition tools don't address?

1. How does the system handle edge cases where requirements conflict with AI-generated content?

1. What is your roadmap for moving beyond the current prototype to a production-ready system?

1. How do you plan to validate that the deterministic checks actually work as claimed in practice?

1. What are the actual constraints on what can be verified deterministically versus what must remain subjective or human-reviewed?