Fresco

AI-powered Division 8 (door, frame, and hardware) takeoff and estimating

Y Combinator-backed (F24) AI company whose current flagship product automates Division 8 (doors, frames, and hardware) takeoffs and estimating for commercial construction, cross-referencing door schedules, floor plans, hardware specs, and elevations to flag discrepancies before a bid goes out. The company's original 2024 product (marketed at fresco-ai.com) was a broader generative-AI documentation copilot that turned superintendents' site-walk videos and photos into reports and punch lists; that jobsite-documentation use case appears in early press coverage but is not featured on the company's current Division 8 marketing site (fresco.build).

Not publicly disclosed as a specific total. Y Combinator (F24 batch) backed; the company site also lists SignalFire, Bessemer Venture Partners, Hand Ventures, Transpose Platform, HomeTeam Ventures, and Flex Capital as investors. Tracxn categorizes the company as seed-stage. Note: some AI-generated search summaries report "$48.2M across 7 rounds" and a "$17.5M Series C" for a company called "Fresco," but this could not be corroborated by any primary source and appears to conflate this 2-person, YC-seed-stage company with a different, unrelated "Fresco" entity — treated as unverified and not used. raisedEarly Stage

Founded 2024 · 70+ companies across the US and Canada use Fresco for Division 8 takeoffs, per the company, on projects ranging from roughly 30-door tenant improvements to 4,000-door hospital renovations; the company also states 3,000+ takeoffs completed. No named customer logos or case studies are published. Y Combinator lists the team at 2 founders; Tracxn categorizes the company as seed-stage.

Website →← All Tools

WHY IT EXISTS

Why This Tool Exists

The Problem

Division 8 (doors, frames, and hardware) is widely considered the most complex takeoff in commercial construction: estimators must manually cross-reference door schedules, floor plans, hardware specifications, and elevations across hundreds of pages to catch every door and assign the correct matching hardware set. A single missed door or misread hardware group — such as an 8-foot door that needs four hinges instead of three — can erase a bid's margin, and on large projects (the company cites hospital renovations with up to 4,000 doors) the manual process can take more than a day per estimator.

The Solution

Fresco's AI reads the door schedule, floor plans, hardware specification, and elevations together, normalizes hardware sets into a structured format, cross-references them against the door schedule, and flags discrepancies (missing doors, mismatched hardware, sizing exceptions) for the estimator to resolve. The output is a validated set of hardware assignments that exports as a structured Excel file mapped to Comsense's import format, or with one click into eMullion's import wizard — carrying hardware sets, component configurations, and pricing without redundant data entry.

USAGE

How You Use It

Delivery Method

SaaSWeb App

Integrations

eMullion (one-click export into eMullion's import wizard)Comsense (structured Excel export mapped to Comsense's import format, not a direct API integration)Procore (listed on the Procore Marketplace; functional integration scope not detailed on the listing)

Disciplines

Construction, Architecture

Workflows

Data & Automation

Project Phases

Bidding/Procurement

Project Types

Commercial, Healthcare

TRANSPARENCY

Data Transparency

Exactly what this tool uses and how

Input

What it needs

Required:Door schedule, Floor plans, Hardware specification, Elevations

Optional:Existing budget or pricing data

Formats:PDF

Output

What you get

Format:Structured Division 8 takeoff with validated door and hardware assignments

Fields:Door counts and specifications, Normalized hardware set assignments, Flagged discrepancies/exceptions (missing doors, mismatched hardware), Component sizing rules (e.g., hinge count driven by door size)

Algorithm

How it works

Model:Proprietary AI the company says is "trained on millions of construction drawings" (specific model architecture/training method not published). The company's own privacy policy discloses that customer input, output, and personal data may be shared with and processed by third-party "AI Service Providers" including Anthropic, OpenAI, AWS, Google Cloud, and Microsoft Azure — suggesting the product is built on top of third-party foundation models rather than a purely from-scratch proprietary model, despite the "proprietary AI" marketing language.

Accuracy:Company claims 99% accuracy on takeoffs and a 70% reduction in estimator time versus manual takeoff. These are self-reported figures; no independent benchmark or methodology has been published.

Privacy

How your data is protected

Retention:Personal information is retained for 12 months following account termination, per the company's privacy policy. De-identified/anonymized data — including takeoff inputs and outputs — may be retained indefinitely for product improvement, AI model training, and internal benchmarking.

Training:Company privacy policy states anonymized, de-identified takeoff data may be used to improve Fresco's AI models and product performance, and in limited sales/demo contexts; raw, identifiable project data is stated not to be shared with third parties in identifiable form.

Compliance:Company states it is "SOC 2 Type 2 certified" on its marketing site and maintains a trust center (trust.fresco-ai.com); the certification report itself is not publicly available and is not referenced in the privacy policy text, No GDPR, ISO 27001, or other certification publicly disclosed

API

Integration

Endpoint:Not publicly documented — no self-serve developer API found

Method:Not publicly specified

USE CASES

Use Cases

·Automating Division 8 (door, frame, and hardware) takeoffs from architectural plans and specifications
·Cross-referencing door schedules, floor plans, hardware specs, and elevations to catch missed doors or mismatched hardware before bidding
·Exporting validated hardware takeoffs into Comsense or eMullion for estimating and pricing
·Estimating Division 8 scope on large, complex projects (from small tenant improvements up to 4,000-door hospital renovations)
·Speeding up bid turnaround for general contractors and Division 8 subcontractors/distributors competing on commercial projects

PRICING

Pricing

Free

A free first takeoff is offered via "Get Your Free Takeoff" on fresco.build; no self-serve free tier for ongoing use is disclosed

Pro

Not publicly listed for the current Division 8 takeoff product — contact vendor for pricing. (Fresco's earlier, separate AI documentation copilot for superintendents was reported in 2024 press coverage at approximately $1,000/site/month with unlimited users, with discounts for multi-month or multi-site commitments; that figure predates the current Division 8 product and its accuracy today is unverified.)

Enterprise

Contact vendor for custom/enterprise pricing

Website →

Research Sources & Data QualityLast verified: 2026-07-16

Data Sources (11)

Verified Data (13)

problem, solution, algorithm-approach, integrations, yearFounded, customerBase, privacy-retention, privacy-training, deliveryMethod, disciplines, projectPhases, input-formats, output-structure

Not Found (6)

fundingTotal-specific-amount, current-pricing, api-documentation, compliance-certification-report, data-hosting-location, algorithm-model-architecture

Our Commitment: We only include verified data from official sources. If information isn't publicly available, we mark it as "Not publicly specified" rather than guessing.

ALTERNATIVES

Similar Tools

🌱

2050 Materials

Data & Automation

Digital information and research platform for the AEC industry to explore, compare and source sustainable building materials with comprehensive sustainability data.

Integrate real-time data on building material climate impacts with API-first solution. Access the industry's largest, highest-quality sustainability database for immediate embodied carbon calculations.

AI platform that transforms natural language questions into interactive maps, charts, and spatial analysis for urban planning and site selection without requiring GIS or coding skills.

View Details →