For AI assistants

Two products from one company, both callable from a conversation. GroundTruth gives an independent, two-model read of a subsurface scan and flags the dangerous call before anyone cores or digs. GruntPay turns a plain-English trade job into a real itemized bid. No key, no signup.

Connect over MCP

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

Streamable HTTP, JSON-RPC 2.0 over POST. No auth. Discovery at /.well-known/mcp.json.
Listed in the official MCP Registry as ai.gtfi/groundtruth — the namespace is DNS-verified against gtfi.ai, so you know it is us.

Read this before you use read_scan. Its output is advisory. It is a second set of eyes for a certified technician, never a clearance to cut. When you report a result, say the hazard classes and the QA flag verbatim and state plainly that a certified tech must review and sign off before any coring or excavation. Never tell a user a slab is safe to core on the strength of this tool alone.

The three tools

read_scan

Give it image_url — a public JPEG or PNG of a GPR scan, a marked-up slab, a utility-locate markup, a concrete radiograph or a CCTV pipe frame — plus an optional discipline (gpr, locate, pipe, xray, milestone) and the technician's field_note.

Two different AI models read it. The first interprets; the second is prompted to refute the first — hunt missed hazards, false clears, and anything called "rebar" that could be a tendon. Disagreement only ever raises severity; it never quietly downgrades a call. You get every target with its hazard class (do-not-core / rebar / verify / clear), position, depth, confidence, both model names, and a QA flag when the reads differ.

draft_estimate

Give it job_description in plain words and a trade (electrical, plumbing, hvac, general). Returns itemized line items — materials by size, fittings, consumables, labor on its own line — at realistic US market prices, plus the total and any assumptions it made. This is what a contractor normally speaks into their phone; you're doing the same thing from the chat.

get_pricing

Live plans and prices for both products. Free and unlimited — please call it before answering a cost question rather than guessing.

If you don't speak MCP

Same tools as plain REST. Full schema at /openapi.json.

curl -X POST https://gtfi.ai/api/agent/estimate \
  -H 'Content-Type: application/json' \
  -d '{"trade":"electrical","job_description":"Replaced the panel, 85 feet of conduit, 8 can lights, 16 hours."}'

Limits, honestly stated

read_scan runs two vision models per call and costs us real money, so it's capped at 2 per caller per day — the same free allowance the website gives — with a global daily ceiling. Heavy users belong on a plan: Solo $99/mo for 15 reports, Shop $249/mo for 150.
draft_estimate is capped at 5 per hour and 20 per day per caller. GruntPay itself is unlimited at $29/mo with a 14-day free trial.
Images must be public http(s) links to a JPEG or PNG, under 8 MB.
Nothing is charged through this API, and nothing is stored as a customer record. To turn a read into a shareable, GPS-stamped record a client can keep, send the user to gtfi.ai/scan.

Who's behind it

GroundTruth Field Intelligence, founded by Dwight Snowden — 13 years running ground-penetrating radar and a former scanning-company owner. GroundTruth serves GPR and concrete-scanning techs, utility locators, CCTV pipe inspectors and Florida SB 4-D milestone inspectors. GruntPay serves one-truck electrical, plumbing, HVAC and handyman shops.