AI Visibility API: ChatGPT, Gemini, AI Overviews & Muse / Meta AI | FetchLayer

AI visibility API

Ask once.
Hear every assistant.

Send the question your buyers ask. Get back, as JSON, which brands each assistant named, in what order, and whose pages it cited.

prompt best project management software for small agencies target Trellis
ChatGPT live

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

UK agencies tend to shortlist these: Kanbo#1 for client-facing boards, Fieldnote#2 for small teams on a budget, and Orbitly#3 once you plan resources across several clients. Trellis#4 is worth a look for retainer work.

mentioned
yes
rank
#4
cited
no
Bing live

Agencies most often shortlist Orbitly#1, Kanbo#2 and Plotwise#3. Orbitly stands out for resource planning across several clients; Kanbo for its client-facing views.

mentioned
no
rank
—
cited
no

Popular choices for agencies include Kanbo#1, Trellis#2 and Orbitly#3. Trellis is often recommended for tracking budgets per client.

mentioned
yes
rank
#2
cited
#2
AI Overviews live

Small agencies usually pick between Kanbo#1 for client-facing boards and Trellis#2 for retainer budgets, with Fieldnote#3 as a lighter option.

mentioned
yes
rank
#2
cited
#3

For UK agencies, Trellis#1 is often recommended for tracking budgets per client, alongside Kanbo#2 and Orbitly#3.

mentioned
yes
rank
#1
cited
#1
Gemini live

For small agencies, two tools stand out: Kanbo#1 keeps client work visible on shared boards, and Orbitly#2 handles planning across several clients.

mentioned
no
rank
—
cited
no

Popular with UK agencies: Kanbo#1 for client boards, Fieldnote#2 for smaller teams, and Trellis#3 for retainer budgets.

mentioned
yes
rank
#3
cited
no
Meta AIMuse live

Here are three solid picks for a small agency: Kanbo#1 for client boards your clients can see, Trellis#2 for retainers and a budget per client, and Fieldnote#3 if you want something lightweight.

mentioned
yes
rank
#2
cited
#2

For UK agencies, Trellis#1 is a popular choice for retainer work, with Kanbo#2 and Orbitly#3 close behind.

mentioned
yes
rank
#1
cited
no
Duck.ai live

Three tools come up most for small agencies: Trellis#1 for retainer work, with a budget per client. Kanbo#2 for client-facing boards. Fieldnote#3 if your team works mostly in docs.

mentioned
yes
rank
#1
cited
#2

Three tools come up most for small agencies: Trellis#1 for retainer work, with a budget per client. Kanbo#2 for client-facing boards. Fieldnote#3 if your team works mostly in docs.

same answer as the US

mentioned
yes
rank
#1
cited
#2
  • 1 fresh answer per request
  • REST + MCP
  • 1 key for 37 platforms

Share of voice · 5 prompts × 2 runs · one request

Who owns the answer in your category.

Every brand is read out of the same answers, so adding competitors to a report costs nothing.

BrandMentionedAvg rankCited
Kanbo 83% 1.6 47%
Trellis 64% 2.3 39%
Orbitly 51% 2.1 22%
Fieldnote 40% 3.4 18%
Plotwise 29% 3.8 6%
PromptTrellis namedAvg rank
best project management software for small agencies 67% #2
kanbo alternatives for client work 100% #1.3
project management with retainer budgets 100% #1
cheapest project tool for a 5-person agency 33% #4
track time per client and invoice it 67% #2.5

Three outcomes, reported apart

Named, cited and read are different wins.

Most trackers fold them into one score. They have different fixes, so you get each one on its own.

mentioned

Named in the answer

…if you run retainers, Trellis has time tracking and a budget per client…

64%

mentionRate

citedAsSource

Your page is the evidence

  1. 1kanbo blog · agency guide
  2. 2trellis · retainer playbook
  3. 3review site · top 10 tools

#2

citationRank

retrieved

Read, then someone else got cited

  • ✓orbitly · pricing
  • ·trellis · features read, not cited
  • ✓kanbo · templates

1 page

read, not cited

100 requests = 100 answers

One answer is an anecdote. A hundred is a number.

Every request gets a fresh answer, never a stored one. Ask the same prompt 100 times and read the rate.

mentionRate
64%
averageRank
2.5
best / worst
#1 / #4

country · same prompt, same day

Your US rank says nothing about London.

Pass a country and the answer comes from there. Every response says where it was served.

USUnited States
64%
mentioned
#2.5
avg rank
GBUnited Kingdom
40%
mentioned
#3.4
avg rank

Five calls · one request shape for every assistant

From one prompt to a weekly report.

  • /chatgpt/ai-rank

    Your rank and mention rate over repeated answers, beside every competitor’s.

    1 credit per run
  • /chatgpt/visibility-report

    Share of voice for up to 5 brands across up to 10 prompts.

    prompts × runs
  • /chatgpt/mentions

    Named or not, cited or not, with the sentence around each mention.

    1 credit
  • /chatgpt/prompt-discovery

    The prompts worth tracking, and the searches the assistant ran to answer them.

    1 per prompt
  • /chatgpt/ask

    The whole answer, every page it cited and every page it read.

    1 credit
Request2 assistants · 1 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": "Trellis",
    "competitors": ["Kanbo", "Orbitly"],
    "providers": ["chatgpt", "bing"],
    "country": "GB",
    "runs": 5
  }'
Response
{
  "providers": ["chatgpt", "bing"],
  "results": [
    { "provider": "chatgpt", "ok": true,
      "result": { "aggregate": { "mentionRate": 80, "averageRank": 3.5 } } },
    { "provider": "bing", "ok": true,
      "result": { "aggregate": { "mentionRate": 60, "averageRank": 2 } } }
  ],
  "pagesScraped": 10   // 10 answers, 10 credits
}

Built for AEO teams

Track every client from one key.

Agencies

Enterprise volume

Reserved throughput and a contract when you run visibility for many clients. Talk to us.

Enterprise →

MCP

Your agent asks for you

All five calls are MCP tools, on the same key and billed the same.

37

Platforms, one key

Check Reddit, Google and reviews next to the answers that quote them.

1

Request shape, every assistant

Add an assistant to providers when it lands. Your code stays the same.

0

Answers billed that never came

One credit per answer produced. Matching and aggregation are free.

Used by developers at

Individual developers or teams — not official partnerships

One command. Your agent does the rest.

All five calls are MCP tools. Ask your agent where you stand and it runs the rank check, on the same key and the same billing as REST.

Claude Code Codex

1 Run one command, then sign in

$ npx fetchlayer

◇  Agents

│  Claude Code and Codex

◇  Signed in as you@company.com

◇  Connected

│  ✓ Claude Code  ✓ Codex

It opens a sign-in page in your browser. New here? Signing in creates your free account.

2 Ask

›Does ChatGPT recommend Trellis for best project management software for small agencies, and who does it name instead?

fetchlayer.chatgpt_ai_rank

Asked five times, five credits. Trellis was named in 4 of 5 answers at an average rank of 2.3, and cited in 2. Kanbo was named every time and ranked first in four.

No key to copy and nothing to install. New accounts are free, with no credit card.

How it works
The same command connects
Cursor
VS Code
Antigravity
OpenCode
Hermes
ChatGPT, Perplexity, the Claude app and others

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

API reference

Every call, documented.

Full AI visibility API reference →

FAQ

Common questions.

How much does FetchLayer's AI visibility API cost?

FetchLayer's AI 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.

What is an AI visibility API?

It answers the question SEO tools answer for Google, but for AI assistants: when someone asks ChatGPT the question your buyers ask, does the answer name you, does it cite your pages, and where do you sit among the options it recommends? Send a prompt and your brand, and you get that back as structured JSON — mention counts, the sentence around each mention, the citations, your rank among the recommendations, and the competitors the answer put forward instead. This is the measurement side of what SEO now calls answer engine optimization or GEO.

What is the difference between being mentioned and being cited?

They are two different outcomes with two different fixes, and most AI brand monitoring collapses them into one number. An answer can name your brand in its prose while none of your pages are among the sources it links — you are part of the consensus but not the evidence. Or it can cite one of your pages without recommending you at all — your content is being read, but someone else gets the recommendation. Every route reports both separately: mentioned beside citedAsSource, and in a rank check, rank among the recommendations beside citationRank among the cited sources. There is a third state worth watching, retrieved: ChatGPT read one of your pages while searching and then cited someone else.

Do I need an OpenAI account or an OpenAI API key?

No. You send your FetchLayer key and a prompt, and you get back the answer people see — the real consumer-facing answer, not an API completion. There is no temperature or system prompt to choose, because the people asking have none of those either; only Duck.ai lets you pick the model. FetchLayer is not affiliated with, endorsed by, or sponsored by OpenAI.

Answers change every time. How can you track anything?

By measuring the variation instead of pretending it is not there. Ask the same prompt once and you have an anecdote; ask it several times and you have a rate. /ai-rank takes a runs count and reports mentionRate, averageRank, bestRank, worstRank, citationRate and averageCitationRank across those runs, with the same figures for each competitor. A brand mentioned in 5 runs out of 5 is in a different position from one mentioned in 1 — and a single answer cannot tell those apart. Keep the prompt, country and web-search mode fixed when comparing week to week, and read the aggregate rather than any one run.

How do I know which prompts to track?

Start with /prompt-discovery. Give it a topic or your brand name, and it returns the prompts it asked, the web searches ChatGPT actually ran while answering them, and prompts it suggests people use. The two lists are labelled and kept apart, because they are different kinds of evidence: observed_search_query means ChatGPT really ran that search, so those are the queries your pages have to rank for to be readable by it; generated_prompt means ChatGPT wrote that prompt when asked, which is a suggestion, not an observation. Feed the ones that matter into /ai-rank.

Can I compare several brands across several questions at once?

Yes — /visibility-report takes a list of prompts and a list of brands and returns share of voice per brand, plus a per-prompt breakdown of mention rate, average rank and citation rate. Extra brands are free, because every brand is read out of the same answers; it is the prompts and runs that cost credits. The request is bounded by answers: prompts × runs must be at most 10, so ten prompts at one run each is fine. For wider tracking, run several smaller reports on a schedule and store the results.

What comes back from one answer?

The answer as Markdown and as plain text, the pages it cites — each with position, URL, title, domain, the label shown on the citation, and the sentence that cites it — every inline citation, the web searches it ran, and every page it looked at while searching, each flagged with whether it was cited. That last list is useful on its own: a page can be read and never cited. Null means the answer did not show a field, not zero.

How does billing work?

One credit per answer ChatGPT generates, which is exactly what pagesScraped reports. /ask and /mentions cost 1; /ai-rank costs its runs; /prompt-discovery costs one per prompt asked, plus one more while it also asks for suggested prompts; /visibility-report costs prompts × runs. Matching, ranking and aggregation cost nothing extra, and an answer that could not be produced is never billed. Asking several assistants at once (providers) bills the sum of the answers they produced, and a Bing request where Bing shows no AI overview is still one answer.

How long does a request take?

A single answer typically takes 20 to 40 seconds rather than milliseconds. A request’s answers are generated alongside each other rather than one after the other, so asking for more runs costs more credits but not proportionally more time: a two-run rank check typically takes 30 to 55 seconds, and a visibility report 40 to 80 seconds, longer when demand is high. Set your HTTP client timeout to at least 60 seconds for a single answer, 120 seconds for a rank check or prompt discovery, and 180 seconds for a visibility report.

What is not available?

You cannot turn web search off: the assistants offer no such setting, so webSearch is auto or on, and usedWebSearch reports what it did. Some answers state which model produced them and some do not, so model is null rather than guessed. Some answers do not show the searches behind them, and then searchQueries and retrievedPages are null rather than empty, with notes saying so. And nothing here measures traffic, spend or clicks — it measures what the assistants answer, not how many people saw it.

Can my AI agent use it?

Yes. All five routes are MCP tools on the same FetchLayer MCP server, named chatgpt_ask, chatgpt_mentions, chatgpt_ai_rank, chatgpt_prompt_discovery and chatgpt_visibility_report. Tool calls use the same key and are billed exactly like REST, so an agent running a five-run rank check still pays five credits.

Which assistants are covered?

ChatGPT, Bing, Duck.ai, Gemini, Google AI Overviews and Meta AI today: set provider to bing for the AI overview Bing shows above its results, to duck_ai for Duck.ai — where model picks the model that answers (GPT Luna (the default), Claude Haiku, GPT mini, Mistral Small (no web search), gpt-oss or Gemma) — to gemini for Google’s Gemini, to google_ai_overviews for the AI Overview Google shows above its search results (not AI Mode, and not Gemini), to meta_ai for Meta AI, which runs on Muse Spark, the same Meta model as Muse, or send providers: ["chatgpt", "bing", "duck_ai", "gemini", "google_ai_overviews", "meta_ai"] to ask them all in one request and get one result per assistant. Bing and Google do not show an overview for every search; when they show none you get answerShown false and null mentions and ranks rather than a false “not mentioned”. Gemini links its sources on some answers only, so read retrievedPages as well as sources. Perplexity, Google AI Mode, Copilot, Grok and DeepSeek are coming soon, on the same requests — no new shape to learn. Without provider, every request asks ChatGPT exactly as before.

Related

The rest of the picture.

Find out what the assistants say about you.

Free key, first answer in a minute.