More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more
    MA

    Measure a Floor Plan Photo Honestly

    1

    The method for turning a photo of a floor plan into measurements you can trust to about one percent — and knowing where not to trust it.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    MA

    Measure a Floor Plan Photo Honestly

    Example session with this skill installed

    Calibrate this imported plan from photo-capture. I have a horizontal width of 12.5m and a vertical depth of 8.2m. Generate the draft record.

    • Read your context and instructions
    • Compiled the measure a floor

    The plan has been processed using two-axis calibration. Perspective warping was applied to the detected paper corners. All elements are currently flagged as verified: false and the export lock is active.

    Import Draft Record

    {
      "provenance": {
        "source": "photograph",
        "calibration": {
          "horizontal_scale": 1.042,
          "vertical_scale": 1.038,
          "method": "two-axis-manual",
          "status": "calibrated"
        },
        "metrics": {
          "mean_confidence": 0.84,
          "wall_count": 14,
          "openings_detected": 6
        }
      },
      "geometry": {
        "status": "UNVERIFIED DRAFT",
        "verified": false,
        "elements": [
          {
            "id": "wall_001",
            "type": "exterior",
            "length_m": 12.502,
            "thickness_mm": 300,
            "confidence": 0.98
          },
          {
            "id": "wall_002",
            "type": "exterior",
            "length_m": 8.195,
            "thickness_mm": 300,
            "confidence": 0.97
          },
          {
            "id": "window_001",
            "parent": "wall_001",
            "width_mm": 1145,
            "confidence": 0.52
          }
        ]
      }
    }
    

    Next steps

    1. Enter Verify Mode to confirm the exterior dimensions (12.5m x 8.2m).
    2. Manually inspect window_001 as it fell below the high-confidence threshold (0.52).
    3. Confirm wall thicknesses to flip the verified flag and unlock export modules.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Software can trace a photographed floor plan in minutes. The hard part is scale: get it wrong and every wall is off by the same percentage, neatly and confidently. This package is the calibration discipline behind my Home Builder AI product — how to set the scale from two known dimensions, how to score each detected wall and opening by confidence, and how to keep cost estimates locked until a person has checked the outside dimensions.

    Scope, stated plainly: this is the method, the measured accuracy data, and the verification design. It is not the Home Builder AI software itself. Inside the package: the skill file, a worked example of an import record (the data a plan carries about its own reliability), and an accuracy-expectations reference with the real test numbers.

    The numbers, from tests run both ways on the same photographed plan: calibrating with a single dimension lost one wall entirely and left the rest off by 3.57% on average, 4.74% at worst. Calibrating two axes — one known width, one known depth — under careful clicking gives roughly ±1% mean error on wall lengths, ±1.5% at worst, the same band professional scan-to-plan services report. A run goes: photograph straight-on with a margin of paper, calibrate both axes, detect walls and openings, then open the result in Verify Mode — badged as an unverified draft, estimates and exports locked — and confirm outside dimensions first, then opening widths, then wall thickness. The honest sentence for any imported plan is: a measured first draft in minutes. Never "exact," never "permit-ready."

    Requires two perpendicular known dimensions for photos. Windows and doors have a higher error margin (25-65mm) than walls. Thickness values are hints only.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 6 days ago

    Requires

    A photo or scan of a floor plan
    Two known dimensions for calibration

    Permissions

    Write Files
    Read Files
    Listed6 days ago

    What's inside

    Frequently Asked Questions

    Unlike basic AI prompts that guess dimensions, this skill requires two-axis (horizontal and vertical) calibration points to correct perspective warp and systematic scale errors, ensuring a mean wall-length error of ±1%.

    The skill exports structured geometry in a schema-valid JSON format designed for seamless integration with CAD, BIM software, and any tool compatible with the `import-record.json` standard.

    Yes, the skill includes a 'Verify Mode' enforcement that locks all downstream exports and estimates until a human reviewer confirms the primary dimensions and clears the 'verified' flag.

    This skill requires at least one horizontal and one vertical known dimension to be marked; if a photo lacks these perpendicular reference points, the skill will reject the file to prevent scale drift.

    You will receive the core computer vision logic, the two-axis calibration engine, the ink-coverage testing module, and the structured schema emission tool for your integration.