PlanRuntime

PlanRuntime gives AI assistants specialized tools and a visual workspace for working with construction plans. Our first use case is aluminum fence and railing takeoff for complex multifamily projects. My father owns both the installation company and the manufacturing company, so one large job creates work across both businesses. Extensive plan sets, repeated railing configurations and technical details require careful interpretation. His estimator cannot keep up with the bids. Depending on the job, installation may involve his own crew or a local subcontractor. He began using an AI assistant to help with takeoffs and now uses it across his businesses. Useful results took substantial coaching about his company. That solved part of the workload; it did not connect the whole process. Checking the work and moving information between plans, workbooks, people and software remain important. PlanRuntime combines a construction MCP server with a shared visual workspace and persistent project records. The design connects source drawings, interpretations, measurements, calculations, review decisions and outputs. An assistant can use specialist plan operations while the owner inspects evidence and makes decisions. Supported previous-takeoff import and company mappings help configure a familiar workbook output. This is structured reuse, not a claim of automatic model training. The broader vision aligns with Muse for Small Business: an assistant that understands the business, prepares work in the background and coordinates connected tools. We envision a bid package becoming an organized project, owner-reviewed takeoff and tailored handoffs for manufacturing and installation partners, with shared files keeping people informed. Those broader connections remain future work. This is a prototype. Current validation includes synthetic company-setup, review and workbook-export workflows; complex-job quantity accuracy and the consumer Muse integration remain open. Unknown values and incomplete review must remain visible. Our goal is practical delegation: make specialized plan work accessible to the owner's assistant while preserving the evidence and judgment needed to use it responsibly.