AEC Spec Validator - Sell Project

AEC Spec Validator is a live B2B SaaS for architecture, engineering and construction teams that need repeatable, auditable validation of technical requirements against BIM/model data. The platform connects the full validation chain: source specification → structured requirement → deterministic rule → BIM evidence → finding → human review → audit output The core compliance path is deterministic rather than AI-driven. Missing or unsupported evidence remains explicitly unknown instead of being guessed or silently treated as compliant. Deterministic BIM validation AEC Spec Validator processes real IFC2x3 and IFC4 models using web-ifc/WASM and supports buildingSMART IDS 1.0 for machine-readable BIM requirements. In addition to focused built-in engineering rules, the native engine includes a generic IFC property and quantity rule system that can validate properties discovered directly in uploaded models across IFC element types. Supported comparisons include: • equals / not equals • minimum / maximum • numeric ranges • in-set / not-in-set • exists / not-exists The No-Code Rule Builder can populate element types, property sets and properties directly from the uploaded IFC model rather than relying only on a fixed rule catalogue. Evidence Assistant The product also includes an evidence-grounded AI assistant for asking questions about project requirements, BIM evidence and validation findings in natural language. Unlike a generic BIM chatbot, the Evidence Assistant does not decide compliance. Every question re-runs the deterministic validation logic on the server. AI responses are checked against known requirement and model-element references before they are shown. Unsupported citations are rejected and the system can fall back to deterministic evidence-based output. AI explains the evidence. It does not decide compliance. This allows engineers and reviewers to ask questions such as: • Which requirements currently fail? • What cannot be determined from this model? • Why did this room or element fail validation? • Which findings require human review? • What evidence supports this result? Core capabilities • Native IFC2x3 / IFC4 model processing • buildingSMART IDS 1.0 parsing and validation • Generic IFC property and quantity validation • Deterministic pass / fail / unknown / not-applicable outcomes • No-code rule authoring • PDF, DOCX, XLSX and CSV specification workflows • Requirement-to-rule-to-evidence traceability • Evidence-grounded AI assistant • Server-side AI citation verification • Human review and waiver workflows • Baseline and regression comparison between revisions • Traceability matrix and coverage metrics • SHA-256 checksum-aware audit packages • CSV and XLSX exports • REST API and CLI access • Webhooks and CI-oriented workflows • Authentication, authorization and Supabase Row Level Security • Team and project access controls • Asynchronous validation jobs • Starter / Professional / Enterprise plan enforcement • Stripe subscription billing • Automated unit, end-to-end and performance testing Auditability instead of a black box AEC Spec Validator is designed so every important decision remains inspectable. A reviewer can follow a requirement from its original source through the validation rule and observed BIM evidence to the resulting finding and final human decision. Unknown information remains unknown, AI cannot override the deterministic engine, and audit exports preserve the evidence behind the result instead of producing only a final compliance percentage. Current product status The product is live on its own custom domain and includes a working Riverside Office demonstration environment with seeded requirements, passes, failures, unknown findings, review decisions, waivers and revision comparison. The application includes production persistence, user registration, project isolation, billing infrastructure, API/CLI workflows and automated QA. The project is currently pre-revenue. Customer acquisition has not yet been the primary focus, so the main opportunity for a buyer is commercial distribution rather than rebuilding the technical foundation. Autodesk integration path The repository also contains a C#/.NET integration-boundary prototype for future native Autodesk workflows. It defines an extractor architecture for eventually connecting Revit or AutoCAD model data to the normalized validation pipeline. This is intentionally described as an integration prototype, not as a finished Revit or AutoCAD plugin. Typical use cases • BIM quality assurance • buildingSMART IDS validation • Design requirement checking • Technical specification compliance • BIM coordination and project QA • Model revision regression checks • Pre-handover verification • Engineering review workflows • Audit-ready validation evidence • Automated BIM quality gates in CI/CD Why the project is interesting Many BIM validation workflows still combine spreadsheets, manual specification reviews, isolated model-checking tools and disco...