ChatGPT Visibility API — Brand Mentions & Rank Tracking | FetchLayer

Is ChatGPT
recommending you?

Brand mentions, ranks and citations in ChatGPT’s answers, as JSON. Ask the question your buyers ask as often as you like, and read the rate.

prompt best project management software for small agencies provider chatgpt target Trellis
ChatGPTanswer

For a small agency juggling client work, these are the strongest picks: Kanbo#1 for boards and client portals in one place, priced per seat. Trellis#2 if you run retainers: time tracking and a budget per client. Fieldnote#3 when your team lives in docs.

mentioned
yes
rank
#2
cited
#1
Same prompt, 5 runsai-rank
  1. run 1 #2
  2. run 2 #1
  3. run 3 #2
  4. run 4 not named
  5. run 5 #3
mentionRate
80%
averageRank
2.0
best / worst
#1/#3
1–5
runs per rank check
4,000
characters per prompt
searchQueries
what it searched
retrievedPages
what it read

Behind the answer

See what ChatGPT searched and read before it answered.

The searches are the queries your pages have to rank for. The pages it read but didn’t cite are the ones closest to winning.

searchQueries

  • best project management tools for agencies 2026
  • agency project management retainer budgets
  • kanbo vs trellis

retrievedPages

  • citedkanbo · agency guide
  • citedtrellis · retainer playbook
  • read, not citedtrellis · pricing
  • citedreview site · top 10 tools

The request · provider chatgpt

Five calls, one key, the same shape for every assistant.

  • /chatgpt/ai-rankRank and mention rate over 1–5 answers
  • /chatgpt/visibility-reportShare of voice across prompts and brands
  • /chatgpt/mentionsNamed or not, cited or not, in one answer
  • /chatgpt/prompt-discoveryThe prompts worth tracking
  • /chatgpt/askThe whole answer and its sources
  • MCPEvery call is also a tool for your agent
Request
$ curl https://api.fetchlayer.dev/chatgpt/ai-rank \
  -H "Authorization: Bearer ss-your-key" \
  -d '{
    "prompt": "best project management software for small agencies",
    "target": { "name": "Trellis", "domains": [ "trellis.example" ] },
    "competitors": [ "Kanbo", "Fieldnote" ],
    "runs": 5,
    "country": "US"
  }'
Response
{
  "aggregate": {
    "runs": 5,
    "mentionRate": 80,
    "averageRank": 2,
    "bestRank": 1,
    "worstRank": 3,
    "citationRate": 40,
    "averageCitationRank": 1.5
  },
  "competitorAggregates": [
    { "name": "Kanbo", "mentionRate": 100, "averageRank": 1.4 },
    { "name": "Fieldnote", "mentionRate": 60, "averageRank": 3 }
  ],
  "runs": [
    {
      "rank": 2,
      "mentioned": true,
      "citationRank": 1,
      "method": "numbered_list",
      "recommendations": [
        { "rank": 1, "name": "Kanbo" },
        { "rank": 2, "name": "Trellis" },
        { "rank": 3, "name": "Fieldnote" }
      ]
    },
    …4 more
  ],
  "pagesScraped": 5
}
// 5 answers, 5 credits

Other assistants

Same request, every assistant.

Start free. Scale when ready.

Try it free, buy credits as you go, or take a monthly plan when your volume is steady.

Free

30 requests
  • Every platform and endpoint
  • REST and MCP
  • No credit card

You pay

$0

30 req × $0
$0
credit card
none
Start free

Pay as you go

never expire
  • No monthly commitment
  • Same access as subscribers
  • ∞ req/min, ∞ API keys

You pay

$1.99per 1,000

1,000 req × $1.99/1k
$1.99
monthly fee
$0
expires
never
Get API key

Monthly plan

Starter

15,000

  • ∞ req/min, ∞ API keys

You pay

$25per month

15,000 req × $1.67/1k
$25/mo
on credits × $1.99/1k
$29.85
you save
$4.85/mo
Get started
  • One page fetched costs one credit.
  • A request that returns no usable data costs 0 credits.
Not sure? Book a 15-min call

Not sure which volume fits?

15 minutes with the founder. Tell us the workload, we tell you what it needs.

Book a call

Guaranteed throughput on the platform you run on

A dedicated platform from $400/mo, plus $500 for each ~21 req/s. Every other platform on the same key. 99.95% uptime, 24-hour support on Slack or Telegram.

Build your plan

FAQ · ChatGPT

Common questions.

Is this the OpenAI API?

No. The OpenAI API gives you a model completion for your own app. This gives you what ChatGPT answers people: the answer, the brands it names, the pages it cites and the searches it ran, as JSON. You don’t need an OpenAI account or key, only a FetchLayer key. FetchLayer is not affiliated with, endorsed by, or sponsored by OpenAI.

Why ask the same prompt more than once?

Because ChatGPT’s answers vary from one ask to the next. One answer tells you what happened once; five tell you how often you are named and where you usually sit. /chatgpt/ai-rank takes runs from 1 to 5 and returns mentionRate, averageRank, bestRank, worstRank, citationRate and averageCitationRank, with the same figures for every competitor. Every run is a fresh answer.

What is the difference between mentioned, cited and retrieved?

Mentioned means the answer names your brand. Cited means one of your pages is among the sources it links. Retrieved means ChatGPT read one of your pages while searching and then cited someone else. They have different fixes, so each is reported on its own: mentioned, citedAsSource and retrieved, and in a rank check, rank beside citationRank.

Which searches did ChatGPT run to answer?

/chatgpt/ask returns searchQueries, the web searches ChatGPT ran for that answer, and retrievedPages, every page it looked at, each flagged with whether it was cited. When an answer does not show its searches, both are null rather than empty, and notes says so.

Does the answer change by country?

Yes. Pass country (US, GB, DE and so on) and the answer comes from there. When a country cannot be served, the answer comes from the default one and notes says so. Keep the country fixed when you compare week to week.

Can I choose the model?

No. There is no model, temperature or system prompt to set, because the people asking ChatGPT have none of those either. When an answer states which model produced it, model reports it; otherwise it is null rather than guessed.

How do I find the prompts worth tracking?

/chatgpt/prompt-discovery takes a topic or a brand and returns the prompts it asked, the searches ChatGPT ran while answering them (observed_search_query) and prompts ChatGPT suggests people use (generated_prompt). The two lists are labelled so an observation is never mistaken for a suggestion.

How much does it cost?

FetchLayer's ChatGPT visibility API returns ChatGPT answers, the brands they mention, the pages they cite and your rank among the recommendations as JSON over REST or MCP. A request is one ChatGPT answer, and each request costs one credit: $1.99 per 1,000 requests pay-as-you-go, or $25 a month for 15,000, with 30 free requests and no card to start. A request that fails is not billed. The same key covers 36 other platforms, and every route is also a tool on FetchLayer's MCP server at https://mcp.fetchlayer.dev, for Claude, ChatGPT, Cursor and other MCP clients.

Find out what ChatGPT says about you.