Nostimates

Getting started

Quickstart

Create your first tracking run and read a distribution back in under five minutes.

This walks through one prompt, three surfaces, ten samples each.

1. Create a run

curl -X POST https://api.nostimates.com/v1/runs \
  -H "Authorization: Bearer $NOSTIMATES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "best CRM for B2B sales",
    "surfaces": ["google_ai_overviews", "chatgpt", "perplexity"],
    "country": "US",
    "language": "en",
    "samples": 10,
    "webhook_url": "https://your-app.example.com/hooks/nostimates"
  }'

Runs are asynchronous. The call returns immediately with a run_id and a status of queued. Collection typically completes in 40–180 seconds depending on surface and sample count.

2. Receive the result

Either poll GET /v1/runs/{run_id} or, preferably, supply a webhook_url and we will POST the completed payload to you once.

{
  "run_id": "run_9f3c21ab",
  "prompt": "best CRM for B2B sales",
  "surface": "google_ai_overviews",
  "country": "US",
  "language": "en",
  "samples": 10,
  "collected_at": "2026-08-03T06:12:44Z",
  "brands": [
    { "name": "HubSpot",   "mention_rate": 0.90, "ci95": { "low": 0.60, "high": 0.98 }, "avg_position": 1.4 },
    { "name": "Salesforce","mention_rate": 0.70, "ci95": { "low": 0.40, "high": 0.89 }, "avg_position": 2.1 },
    { "name": "Pipedrive", "mention_rate": 0.30, "ci95": { "low": 0.11, "high": 0.60 }, "avg_position": 3.8 }
  ],
  "citations": [
    { "url": "https://www.g2.com/categories/crm", "position": 1, "sample_frequency": 0.90 },
    { "url": "https://www.hubspot.com/products/crm", "position": 2, "sample_frequency": 0.70 }
  ],
  "extraction": { "success": true, "parser_version": "aio-2026.07.3" }
}

3. Read the numbers correctly

  • mention_rate is the share of the n samples in which the brand appeared at all.
  • ci95 is the Wilson score interval for that rate at the sample count you paid for.
  • avg_position is the mean ordinal position of the brand's first mention within the answer.
  • sample_frequency on a citation is the share of the n samples that included that exact URL.

Do not report a single sample

At n=1 every mention_rate is 0 or 1 and the interval spans almost the entire range. n=10 is the lowest count we recommend putting in front of a client, and even there the interval around a 30% share is roughly ±25 points.