# GroundTruth — subsurface scan QA & trade estimating [Health: Active]

**Category:** 🏭 Industrial & IoT  
**Repository:** https://github.com/Dranos-Bot/groundtruth-mcp  
**GitHub Stars:** 0  
**Views:** 2  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/groundtruth-subsurface-scan-qa-trade-estimating-2

## Description
Two independent AI models read a GPR, locate or pipe scan and flag the do-not-core call.

## Tools
Capabilities this server exposes over MCP:

- **read_scan** — Get an independent AI second opinion on a subsurface scan before anyone cores or digs. Send a public image URL of a GPR scan, a marked-up concrete slab photo, a utility-locate markup, a concrete radiograph, or a CCTV pipe frame. TWO different AI models read it independently — the second is prompted to refute the first — and you get back every located target with its hazard class (do-not-core / rebar / verify / clear), plus a QA flag when the two reads disagree. Advisory only: a certified technician must review and sign off before any cutting.
- **draft_estimate** — Turn a plain-English description of a trade job into a professional itemized bid — real materials by size, fittings, consumables and labor on its own line, at realistic US market prices. Built for one-truck electrical, plumbing, HVAC and general-contracting shops. This is GruntPay: the contractor normally speaks this into their phone. Returns the line items plus a link to brand it and send it to the customer.
- **get_pricing** — Current GroundTruth and GruntPay plans and prices. Free to call — use it before answering any cost question rather than guessing.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "groundtruth-subsurface-scan-qa-trade-estimating": {
    "url": "https://gtfi.ai/mcp"
  }
}
```

## Documentation

## What GroundTruth — subsurface scan QA & trade estimating MCP server does

GroundTruth — subsurface scan QA & trade estimating MCP server exposes a hosted MCP interface for two related workflows: reviewing subsurface scan imagery and preparing trade estimates. The scan workflow accepts a public JPEG or PNG URL and supports `gpr`, `locate`, `pipe`, `xray`, and `milestone` disciplines. The estimating workflow accepts a plain-language description of completed or planned work for electrical, plumbing, HVAC, or general trade jobs.

Scan results identify located targets and include a hazard class, position, depth, confidence, the names of both models, and a QA flag when the interpretations differ. Hazard classes include critical do-not-core findings, rebar, verify, and clear. Estimate results are itemized with materials by size, fittings, consumables, labor as a separate line, and stated realistic US market prices.

## How it works

The hosted service is available at `https://gtfi.ai/mcp` using streamable HTTP and JSON-RPC 2.0 over POST. A client sends MCP tool calls to that endpoint; the repository provides the public interface contract, registry information, documentation, and an example rather than the server implementation.

`read_scan` sends the image through two different vision models. One produces an interpretation, while the other is prompted to challenge it by looking for missed hazards, false clearances, and possible post-tension tendons that might otherwise be labeled as rebar. A disagreement can increase a hazard severity, but does not silently remove a hazard. An optional field note provides site context and cannot override a hazard visible in the image.

`draft_estimate` interprets the supplied job description and returns an itemized bid. `get_pricing` reports the current plans for the two products and should be called before answering questions about cost or pricing.

## Setup and configuration

No local installation and no API key are required. Add the hosted endpoint to an MCP client with a configuration such as:

```json
{
  "mcpServers": {
    "groundtruth": {
      "type": "http",
      "url": "https://gtfi.ai/mcp"
    }
  }
}
```

The service also publishes discovery metadata at `https://gtfi.ai/.well-known/mcp.json` and an OpenAPI representation at `https://gtfi.ai/openapi.json`. `read_scan` requires a public HTTPS image URL pointing to a JPEG or PNG no larger than 8 MB. Its discipline defaults to `gpr`; `field_note` is optional. `draft_estimate` requires `job_description`, while `trade` is optional.

## Tools and capabilities

- `read_scan`: Performs an independent two-model review of a subsurface image and returns targets, hazard classifications, locations, depths, confidence, model names, and disagreement flags.
- `draft_estimate`: Builds a trade bid from a natural-language job description, covering material quantities and sizes, fittings, consumables, labor, and US market pricing.
- `get_pricing`: Returns current plans for both products, including the available free and unlimited offering described by the service.

## Limitations and notes

GroundTruth — subsurface scan QA & trade estimating MCP server provides advisory output, not a clearance to cut, core, or dig. A certified technician must review and sign off before coring or excavation, and scan hazard classes and QA flags should be reported as returned. The tool alone must never be used to declare a slab safe.

Usage limits apply. `read_scan` allows two calls per caller per day, subject to a global daily ceiling. `draft_estimate` allows five calls per hour and 20 per day per caller. Nothing is charged through this API, and it does not create a customer record. A shareable, GPS-stamped record is available through the separate `gtfi.ai/scan` site. The hosted server implementation is not open source; the MIT license applies to the repository's example code and manifest.

_Full upstream README: https://allmcps.com/mcp/groundtruth-subsurface-scan-qa-trade-estimating-2/readme_