Mass Timber Panelization & Quoting Tool

DEVELOPER BRIEF — FOUNDING ENGINEER SCOPE Panelization Engine & Quoting Tool Mori · Mass Timber Design-to-Fabrication · Pre-Seed Stage Engagement at a Glance Company Context Mori is a pre-seed mass timber design-to-fabrication startup building a model-to-machine software pipeline — a system that converts architectural designs into fabrication-ready outputs for CLT systems. The core problem: getting from a design to a fabrication-ready quote and panel layout today takes weeks of back-and-forth between architects, structural engineers, and suppliers. Mori is building that compression into software — taking it from months to minutes. What We're Building A web-based tool (or well-structured internal app) that takes an architectural model file as input and produces a panelized mass timber layout, a connection takeoff, and a priced quote — without requiring manual back-and-forth with suppliers. This is v1 of what will become Mori's core costing and estimating module. The data model and architecture decisions made here are platform decisions, not throwaway code. Functional Specification Module 1 — Model Ingestion Accept and parse architectural model files in at least one of the following formats for v1, with a clear path to adding the others: IFC (Industry Foundation Classes) — primary target; open standard, widest compatibility Rhino / .3dm — common in design-forward residential and hospitality projects Revit / .rvt or .ifc export — standard in larger multifamily and commercial From the model, extract: Floor plate geometry per level (bounding extents, overall footprint) Structural grid or bay grid (column lines / bearing wall locations) Floor-to-floor heights per level Span direction indicators where present Openings (stairs, shafts, penetrations) that affect panel layout Module 2 — Panelization Engine Decompose the extracted floor plate into a panel layout using rules defined by a fractional structural advisor (delivered as a structured rule document before build starts). The engine will: Lay out CLT panels respecting Sterling's max panel dimensions and weight limits Apply span direction rules (strong vs. weak axis) and ply selection logic Handle openings — trim panels around stairs, shafts, and large penetrations Select CLT ply configuration from Sterling's catalog based on span and load type Output: panel count by type, dimensions, and area per panel — the takeoff table Rules will be provided as a structured input (table / JSON config) — the engineer encodes them, not derives them. The structural advisor is a separate parallel engagement. Module 3 — Connection Selection + BOM Using the panelization output and the connection library (also provided by the structural advisor), the tool selects appropriate connection types and generates a hardware BOM: Evaluate each panel interface (panel-to-panel, panel-to-beam, panel-to-wall, panel-to-foundation) against the connection rule set Assign connection type per interface (e.g., plywood spline, LVL spline, concealed bracket, surface bracket, hold-down) Look up fastener specs per connection type (e.g., SDS screw diameter and length, Simpson bracket model) Calculate quantity per interface based on linear feet / panel count formulas from the rule set Output: connection BOM — item, spec, unit, quantity — suitable for hardware pricing and eventual shop drawing generation Module 4 — Quoting Engine Combine panelization and connection outputs with configurable rate cards to produce a structured cost estimate: Material costs CLT panels: Sterling rate ($16–18/SF for 5-ply; adjustable in config) × panel area Connection hardware: BOM quantities × unit pricing from a hardware price list (editable config) Freight Configurable $/truck rate by origin (Sterling: ~$1,200/truck to New England; adjustable per project region) Truck capacity based on panel weight and count Other cost lines Fabrication / finish: configurable $/SF adder Install labor: configurable $/SF or $/panel adder Contingency: percentage buffer (configurable, recommended 10–15% for v1 quotes) Margin: configurable percentage Output Summary: total cost, $/SF (hard cost), and cost by category Detailed BOM exportable to CSV Quote summary exportable to PDF — clean enough to send to a client or architect All rate card inputs editable via a config file or simple admin UI — not hardcoded Milestones & Timeline Technical Notes These are preferences and constraints, not mandates. If you have a strong reason to deviate, make the case — architecture decisions are a conversation. Preferred libraries: IFC parsing: Python: IfcOpenShell (mature, widely used in AEC) TypeScript/Node: web-ifc or ifc.js (lighter but less complete) Shapely (Python) or equivalent for 2D floor plate operations — clipping, union, panel tiling Geometry: Python (FastAPI) or TypeScript (Node/Express) — founder has no strong preference Backend: Simple, functional UI — this is an internal tool for v1, not a polished product. React or plain HTML+JS both fine. Frontend: Rate cards and structural rules in JSON or YAML — editable without touching code Config: Build the panelization and BOM schema so it can extend to shop drawings and CNC file output later. This is the platform foundation. Data model: Every rate, rule threshold, and material spec lives in config — not in the codebase No hardcoded values: Out of Scope for v1 Why This Problem Is Hard (and Interesting) This is not a CRUD app. The interesting engineering problems are: IFC geometry extraction is genuinely messy — real architect files have inconsistent layer names, broken geometry, mixed units, and missing structural grids. Building a parser that handles this gracefully is a real problem. Panelization is a constrained layout optimization problem — packing panels efficiently within a floor plate while respecting supplier constraints, openings, and span direction is non-trivial. The data model is platform-critical — the schema you design here becomes the substrate for CNC output, shop drawings, and eventually the full model-to-machine pipeline. It matters. This is the first time this workflow has been automated for mass timber at this level of specificity — there is no off-the-shelf library to reach for. Engagement Terms Non-negotiable regardless of engagement structure: Work-for-hire agreement — all code, data models, schemas, and documentation produced during this engagement are owned by Mori Full IP assignment — no carve-outs for pre-existing tools used solely for this build NDA — standard mutual, covering Mori's supplier relationships, pilot clients, and technical architecture No use of Mori outputs to train external models or contribute to open-source projects without written consent How to Engage If this scope is a fit, reach out with: A brief note on your experience with IFC parsing, computational geometry, or AEC tooling — or adjacent domains (point clouds, BIM, CAD/CAM, fabrication software) A project or repo that shows how you approach a messy, real-world data problem Your rate or equity expectations, and availability to start