× Kanopi Internal data room · Working draft · June 2026
Working draft. This data room is the live working surface for the Kanopi productization read. Everything here is provisional and gets updated as calibration work and platform decisions land.

What Kanopi is

Kanopi is a construction estimating platform built around one idea: a builder's own cost history is a better database than any generic cost book.

Two coupled assets make it work. The first is a multi-track rate library calibrated from real projects the practice has bid and built, segmented by building type so residential rates never muddle commercial ones. The second is an autonomous takeoff engine that reads a plan PDF with vision models, cross-checks its own quantity counts, prices the work against the library, and assembles a complete, confidence-tagged bid in two to four minutes of wall time.

Neither asset is a demo. The engine is the live bidding tool inside the LÏEF and Strata construction practice today; the library prices real work every week.

Where it came from

Kanopi was not designed as a software product first. It accreted inside a working estimating practice: every bid the practice produced fed rates, ratios, and sanity checks back into a shared library, and the automation grew around the parts of the workflow that kept repeating. The estimating discipline came before the code, which is the reverse of how most construction software gets built and, in our read, the reason the numbers hold up.

In 2026 the platform widened from estimating alone to the full preconstruction chain. Architecture feeds design, design feeds takeoff, takeoff feeds the bid. Each phase produces a structured output the next phase consumes, and a broken upstream seam blocks the downstream phase by design. Pricing is never produced without verified quantities.

Asset one · Rate library

Calibrated, segmented, source-tagged

160+ rates across six tracks (custom residential through multifamily and hospitality). Every rate carries its source project, its measurement basis, a confidence tag, and a ±10% variance band against the track median.

Asset two · Takeoff engine

Autonomous, cross-verified, auditable

Vision models read the plan set; counts are verified across independent sources before any pricing happens. Three fully autonomous bids have been produced end to end, at roughly a dollar of model spend each.

Current state, honestly

SurfaceStateNote
EngineLiveProducing real bids for the practice; the calibration anchor landed within 7.1% of actual built cost.
Rate libraryLive, single-anchorOne track has a full built-and-firm-cost anchor; others run provisional or pending anchors. Deepening this is the top calibration priority.
Tracker appLive internallyThe bid pipeline and project tracker runs the practice's work; roughly 70% of the way to a sellable multi-tenant app (upload, auth isolation, billing pending).
Design capabilityProven internally onlyAI agent teams producing permit-ready plan sets under human orchestration, demonstrated on one of our own buildings. Roadmap, not a shipped product.
Brand and entityPlaceholderWorking wordmark; real brand package and legal entity are open workstreams.
Public postureLockedThrough 2026 the framing is integration roadmap, never public launch. The product is integration-led, not self-serve-led.

Why we think it matters

Every tier of the estimating-software market, from enterprise suites to entry-level tools, runs on generic cost databases calibrated to no one. A contractor pays thousands a year to look up the same number every competitor in their region looks up. Kanopi inverts that: the operator's own history is the database, and the autonomous takeoff makes using it faster than the lookup. The Market section carries the full competitive read.

How this data room is organized

The Engine explains the rate library and takeoff pipeline in plain English. Platform covers what productization looks like and who delivers it. Market is the competitive landscape. Roadmap holds the 2026 posture, the engineering runway, and the structural decision in front of us.