Building Code Compliance Checker

Part of the worlds-biggest-software-project initiative.

AI-assisted plan review that checks architectural and engineering drawings against building codes, zoning ordinances, and accessibility standards -- cutting permit review cycles from weeks to hours.

Building Code Compliance Checker uses computer vision and NLP to ingest PDF and DWG drawing sets, extract dimensions, annotations, and symbols, then cross-reference them against ICC, NFPA, ADA, Fair Housing Act, and local code amendments. It generates structured correction lists with code citations, severity ratings, and remediation guidance. The tool serves both government building departments looking to accelerate permit reviews and private design firms that want to catch non-compliant conditions before submission.

Why Building Code Compliance Checker?

Permit review is a persistent bottleneck. US jurisdictions routinely take 4-12 weeks per permit review round, and the construction industry loses an estimated $31 billion annually to rework caused in part by unclear or non-compliant submissions.
Every incumbent is proprietary SaaS. CodeComply, CivCheck, PlanCheckPro.AI, Blitz AI, Archistar, Ichi, and UpCodes Copilot are all closed-source commercial platforms. There is no open-source alternative for building code compliance checking.
Multi-jurisdictional configuration is painful. No existing tool offers straightforward self-service management of local code amendments and zoning overlays. Jurisdictions must rely on vendors to update code libraries.
Integration gaps between design and permitting. Most tools integrate with permit management systems but lack deep connections to architect/engineer workflows (Revit, BIM, AutoCAD). Ichi is the only product making progress here.
Underserved use cases remain. Construction-phase compliance checking, appeals/variance documentation, and specialised domains like stormwater, energy codes, and green building standards are largely unaddressed by current solutions.

Key Features

Plan Ingestion and Drawing Analysis

Upload PDF or DWG drawing sets with AI-driven sheet classification (floor plan, elevation, site plan, details)
Computer vision extraction of walls, corridors, stairways, dimensions, annotations, and symbols
Support for varied PDF quality including hand-drawn plans
Cross-sheet consistency validation and completeness checking
Plan quality assessment predicting likely review cycle count

Code Compliance Checking

Automated compliance analysis against ICC/IBC, NFPA, ADA, FHA, and local amendments
Multi-discipline review covering building, fire, accessibility, plumbing, and structural codes
Jurisdiction-specific configuration for code versions, local amendments, and zoning overlays
Accuracy target of 95%+ on compliance decisions
Confidence scoring for edge cases requiring human judgement

Structured Review Output

Itemised correction lists with exact code citations and severity ratings
Specific remediation guidance and suggested corrective actions
Clear pass/fail presentation of compliance test results
Ready-to-share compliance reports for applicants and reviewers
Decision audit trail for liability, appeals, and review transparency

Reviewer and Applicant Workflows

Guided review interface walking human reviewers through required checks systematically
Pre-submission checking tool for architects and engineers to reduce resubmittal rounds
Human-in-the-loop design preserving final decision authority with reviewers
Multi-revision tracking with automatic change highlighting
Team collaboration and workflow management
Mobile reviewer interface for field-based inspections

Integration and Interoperability

Integration with major permit management systems (Accela, Tyler EnerGov, OpenGov)
BIM and CAD tool integration (Revit, AutoCAD export)
GIS integration for spatial context in review workflows
API-driven architecture for custom integrations
Project context memory carrying jurisdiction and document context across sessions

AI-Native Advantage

Current incumbents use AI primarily for drawing element extraction and rule-based code matching. An AI-native open-source approach can go further: NLP-based interpretation of ambiguous, context-dependent building code language where multiple readings are possible; automated generation of remediation suggestions and design workarounds rather than just flagging violations; intelligent RFI and clarification question generation that detects what is unclear in a submission; and historical variance and appeal pattern analysis using jurisdiction data to predict likely outcomes. These capabilities move the tool from a compliance checker into a compliance advisor.

Tech Stack & Deployment

Drawing analysis: Computer vision and NLP pipelines for element extraction and code text interpretation
AI architecture: Multi-modal AI combining document understanding with regulatory knowledge bases
Code management: Jurisdictional code update pipelines as codes are adopted and amended
Deployment modes: Self-hosted for government departments with data sovereignty requirements, cloud-hosted for design firms
Drawing formats: PDF and DWG ingestion
Integrations: REST APIs for permit management systems (Accela, Tyler, OpenGov), BIM/CAD tool connectors (Revit, AutoCAD), and GIS systems

Market Context

The US construction permitting market is substantial, with the industry losing an estimated $31 billion annually to rework. All current AI-native plan review platforms (CodeComply, CivCheck, PlanCheckPro.AI, Blitz AI) are proprietary SaaS offerings. Monetisation models include government department subscriptions (per permit or per jurisdiction), private firm subscriptions (per project or seat), and marketplace models where applicants pay for pre-submission checks. Florida's 2025 House Bill 683 authorising licensed private providers to use automated plan review systems signals growing regulatory acceptance. Primary buyers are municipal building departments facing reviewer shortages and private architecture/engineering firms seeking to reduce correction rounds.

Project Status

This project is in the research and specification phase. Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.