Skip to content
← Directory

mcp · Global

GeomVerity

建筑机电CAD语义解析与可追溯自动化框架:确定性几何、受约束AI、专业规则与工程证据。

Feature or claim this listing

Feature this listing — founding Featured slots are 50% off for 6 months

Represent GeomVerity? Send a claim for review so we can discuss listing updates. Claims do not automatically grant ownership or verification.

Explore paid sponsorship →

What's in it

Data offered docs-verified
CAD semantic objects with source references, built from CAD primitives · rule checks (four plant-room condition rules, plus a draft example) · drawing-revision change analysis (object changes, partition relationships, impact candidates, prior confirmations) · lamp layout candidates along walls, avoiding doors · LPD and utilisation-factor illuminance estimate for closed rooms · traceable decision records with evidence-pack hash verification
Formats docs-verified
JSON · DXF · HTML · Markdown · MCP
Access docs-verified
Open source, licence MPL-2.0. Local Python install (3.11+); the demo needs no AutoCAD or API key. Scanning DWG files requires the CAD plugin. Live AI features may need a key (not stated in docs).

Useful for

Use cases llm-inferred
Revision comparison of building-services CAD layouts to find changed objects and check whether prior engineer confirmations still hold · Auditable record of design-assist decisions, with inputs, rules, model use and sign-off kept together · Rule checking of plant-room or similar room conditions against extracted drawing data
Audience llm-inferred
building services (MEP) engineers · BIM and CAD coordinators · design reviewers and checkers · software developers building CAD or BIM tooling

Limitations

Constraints docs-verified
Tested only with the Tianzheng CAD plugin; other CAD sources are vision only · Docs are in Chinese and the rules are China-oriented building MEP, not UK water, civils or environmental data · Cross-floor, room-name misalignment, exit-eligibility and specialist-measure data may be missing · Curve sampling IsExact flag still needs correction · Lamp layouts are candidates; real luminaire parameters must be checked · Full demo needs about 8GB RAM; a 512MB sandbox was insufficient · Professional conclusions need supporting material and engineer sign-off
Licence
MPL-2.0
Authentication
Unknown — review source docs
GitHub stars
1
Last commit
15/10/2026
Last successful page check
14/10/2026
Page health
healthy
Package
Not recorded
Verification
Not independently verified

Auto-extracted, unverified fields: tools, auth. Check upstream documentation.

Provenance & evidence

Daily discovery 2026-10-15: GitHub search; verified via GitHub API. Tools/install/auth auto-extracted from README.

Metadata observed 16/10/2026 · daily-2026-10-15 (github-search)