Skip to main content

DocsREST API and MCP server

GTM Scoop API quickstart

Endpoints, auth and live samples for Audience Builder, Connect Data and CRM sync.

Base URLhttps://us-farmvision-new-api.onrender.com

Get started

Quickstart

From a new API key to a live audience count, then AI-drafted filters, in four steps.
  1. 1

    Create an API key

    Press Get API key, enter your company email (phone is optional) and paste the key we email you. Once signed in, your API key (it starts with llmcp_) is created under Profile → MCP & API keys and shown once, so store it somewhere safe.

  2. 2

    Count an audience

    Send the key as a Bearer token and POST a filter set to the company search count.

    Request
    curl -X POST "https://us-farmvision-new-api.onrender.com/api/v1/business-search/count" \
      -H "Authorization: Bearer YOUR_API_KEY" \
      -H "Content-Type: application/json" \
      -d '{"filters":{"states":["Texas"],"employeeBuckets":["50-99"]}}'
  3. 3

    Read the response

    Every response has a success flag. The count sits in data.count.

    Response · 200 OK
    {
      "success": true,
      "data": {
        "count": 38412,
        "executionTime": 642
      }
    }
  4. 4

    Let AI draft the filters

    Describe the audience in plain language, up to 2,000 characters. /ai-criteria returns the filters as a list: review them, then send the list unchanged as filters to /count. Anything it could not turn into a filter comes back in unmatched.

    Request
    curl -X POST "https://us-farmvision-new-api.onrender.com/api/v1/business-search/ai-criteria" \
      -H "Authorization: Bearer YOUR_API_KEY" \
      -H "Content-Type: application/json" \
      -d '{"text":"Corporations in Texas with more than 50 employees"}'
    Response · 200 OK
    {
      "success": true,
      "data": {
        "filters": [
          {
            "id": "ai-state-1793599200000-0",
            "filterId": "state",
            "values": [
              "Texas"
            ],
            "exclude": false
          },
          {
            "id": "ai-minimum-employees-1793599200000-1",
            "filterId": "minimum-employees",
            "values": [
              "more",
              "50"
            ],
            "exclude": false
          },
          {
            "id": "ai-legal-status-1793599200000-2",
            "filterId": "legal-status",
            "values": [
              "3"
            ],
            "exclude": false
          }
        ],
        "unmatched": "",
        "usage": {
          "calls": 1,
          "input_tokens": 3488,
          "output_tokens": 301
        }
      }
    }

Next: ask your AI agent

The MCP server runs the same counts from Claude, ChatGPT, Claude Code and Cursor.

Open playground

What you need

  • A work email address to sign in with.
  • cURL, Python 3 with the requests package, or Node.js 18 or later (fetch is built in). The samples above switch between these and more languages.
  • The base URL, https://us-farmvision-new-api.onrender.com. Every path in these docs is relative to it.

How many calls a key can make

Every request made with an API key counts toward its daily limit, which follows your plan and resets at midnight (server time). The per-minute limit below applies to the audience search routes and the MCP server; every other route allows 300 requests a minute per account. GET /api/v1/gtmscoop/api-keys/entitlement returns your tier and both limits, and the playground runs the same requests in your browser if you want to try a filter before you write code.

PlanCalls a day (every route)Calls a minute (audience search and MCP)
Free1020
Starter3060
Growth50150

What a count counts

By default /count returns dedupe: the matching contact records, with duplicates removed. Send countBy: "company" for distinct companies, or contacts, companyName or executives for the other measures. Send countBy as an array to get several measures for the same filters in one call, in data.counts (data.count still holds the first one). To compare segments, make one call per segment.

filters is optional: leave it out to count the whole dataset. Named filters such as states and employeeBuckets are combined with AND, while a size bucket, a range and a comparison on the same measure are combined with OR. The Company Data Search reference lists every filter and the values it accepts.

If the first call fails

StatusWhat it meansWhat to do
401"Not authorized, no token": the Authorization header is missing.Send Authorization: Bearer llmcp_… on every request.
401"Invalid or revoked API key": the key was mistyped, replaced or revoked.Create a new key under Profile → MCP & API keys.
400A filter name or value is wrong. The body names the field at fault.Check the filter against the reference, or let /ai-criteria write it.
429You reached a per-minute or daily limit. Every 429 carries Retry-After (seconds); every response carries RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset.Wait the Retry-After seconds. A daily limit clears at 00:00 UTC; a larger plan raises it.
503Search is busy: the request waited 20 seconds in the server's queue.Retry in a few seconds.
cURL · what a missing key looks like
curl -s -i "https://us-farmvision-new-api.onrender.com/api/v1/gtmscoop/api-keys/entitlement"

# No Authorization header, so:
# HTTP/1.1 401 Unauthorized
# {"success":false,"message":"Not authorized, no token"}

Where to go next

Cookies on GTM Scoop

Essential cookies keep you signed in and the site secure, and are always on. Optional cookies (Google Analytics) help us understand how the site is used. They are on by default, except in the EU/EEA, the UK and Switzerland, where they stay off until you choose Accept All Cookies; choose Essential Cookies Only to turn them off. We do not use advertising cookies and we do not sell or share your personal information. The site works the same whichever you choose. We keep a record of your choice with the date, your IP address and the page you were on. Cookie Policy · Privacy Policy