AI Citations API: list and classify every URL AI engines cite

The AI Citations API lists the pages AI engines cite in their answers, with the domain, the kind of source and whether the page itself names the brand. It shows which sites feed the answers in a category and which of them never mention you, prompt by prompt and engine by engine.

Updated

  • ChatGPT
  • Perplexity
  • Google AI Mode
  • Claude
  • Microsoft Copilot
  • Grok

Playground

4 of the 6 pages Perplexity cites here never name the brand

The AI Citations API returns every page an engine cited for the prompts in scope. In this sample, Perplexity cites six pages for the sample prompt and only two of them, a review site and the brand's own page, mention Loopwork. The other four shape the answer without naming the brand, which makes them the first sources to work on.

POST /v1/sources/domains

prompt

best project management tool for creative agencies

models[]

Pages cited
6
Pages naming the brand
2 of 6
curl -X POST https://api.apiforgeo.com/v1/sources/domains \
  -H "Authorization: Bearer $API_FOR_GEO_KEY" \
  -H "x-project-id: $PROJECT_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "best project management tool for creative agencies",
    "models": [
      "perplexity"
    ],
    "country": "US"
  }'
200 OK · 6 pages
DomainCategoryNames the brand
review-site.com review yes
loopwork.io brand yes
pm-weekly.com editorial no
agencyforum.net forum no
studio-ops.net editorial no
designers-guild.org directory no
Fields returned per cited page by POST /v1/sources/domains
FieldTypeDescription
urlstringThe cited page, as the engine linked it.
domainstringThe domain the cited page belongs to.
categorystringThe kind of source, such as review or brand.
mentionsBrandbooleanWhether the cited page itself names the brand.

Get started

Four setup steps lead to the first list of cited pages

The AI Citations API reads the sources stored with each run, so the first list arrives after four steps: create the project, generate a key in the app, import the prompts and run them. Every enriched run records the cited URLs, and reading them back is not billed.

  1. Step 1

    Create a project

    Create an organisation and a project, one project per brand or per client.

  2. Step 2

    Generate a key

    Generate an API key in the app; keys are managed there, never through the API.

  3. Step 3

    Import your prompts

    Import your prompts and tags with a single POST /v1/queries.

  4. Step 4

    Run and read

    Trigger an execution, poll the job, then read any analytics endpoint.

Comparison

Only 1 of 4 approaches returns classified citations as data

The AI Citations API returns each citation with its domain and category as structured data. An AI scraping API returns the URLs only, a GEO dashboard tool shows the classification inside its interface, and an in-house build has to fetch, parse and classify every source itself, engine by engine.

3 of the 15 capabilities on the full comparison
Capability API for GEOAI scraping APIGEO dashboard toolIn-house build
Citations with domain + category YesURLs onlyIn the UIBuild it
Raw answer text YesYesSometimesBuild it
Programmatic access to everything YesYesExport onlyYes

Pricing

AI Citations API pricing runs from $0.013 to $0.007 per credit

The AI Citations API reads citations captured during normal runs, so the list itself is not billed. Credits pay for the executions, one per prompt, per model, per locale, and every enrichment layer, citations included, comes with that one credit. The tier follows the month's total volume.

0 – 50k / month
$0.013
per credit
50k – 250k / month
$0.011
per credit
250k – 1M / month
$0.009
per credit
1M+ / month
$0.007
per credit

Price per credit, set by the month's total volume.

FAQ

AI citations, answered

What does one call return?

The AI Citations API returns one object per cited page in the scope you set: the URL as the engine linked it, its domain, its source category and whether the page itself names the brand. The scope works like every analytics read: prompt, models, country and a date range.

Can I find the cited pages that leave my brand out?

The AI Citations API flags every cited page with mentionsBrand, so the pages that shape answers without naming the brand are marked in the response. Those pages are the outreach and content targets: the engine already trusts them for the category, and they do not mention you yet.

How does it relate to true reach?

The AI Citations API gives the inputs of true reach, the share of cited pages that also mention the brand. Counting the pages with mentionsBrand set to true and dividing by all cited pages in scope gives the figure for any prompt, engine or market you read.

Are citations billed separately?

The AI Citations API bills nothing on top of the runs. Each enriched run records the cited URLs with its credit, and every read of the citation list afterwards is unmetered, so a weekly sources report costs the executions that produced it and nothing more.

Can I resell citation data?

The AI Citations API allows resale and white-label use like every family. The payload carries no branding and each client can sit in an isolated project, so an agency can deliver source audits under its own name and a platform can add a citations tab to its product.

List the pages AI engines cite for your prompts