Ai Citation Finder

Find which websites influence AI answers by extracting and ranking the citations used by ChatGPT, Gemini, Claude, and Perplexity.

Overview

For every question you provide, this Actor asks each selected AI answer engine, extracts every citation and URL from the answer, normalizes domains, and counts frequency. It then ranks the most influential sources overall and per category (review sites, blogs, communities, forums, social, government, documentation, news).

Architecture

Identical shared library layout as the rest of the AI Actor ecosystem:

src/
├── main.ts                 # Actor.main() entry point
├── types.ts                # Input/record/summary types
├── config/index.ts         # Input validation → strict config
├── analysis/index.ts       # citation extraction + per-domain aggregation
├── report/index.ts         # Final ranked markdown report
└── shared/                 # Reusable library (adapters, engine, parser, citations, …)

Pipeline: validate → build tasks → collect answers (Playwright adapters) → extract & normalize citations → count per (question, domain) → rank by category.

Input

Field	Type	Default	Description
questions	string[] (required)	—	Questions asked on every platform
platforms	string[]	all four	ChatGPT, Gemini, Claude, Perplexity
topN	integer	20	Domains per ranked category
headless	boolean	true	Headless browser
screenshotOnFailure	boolean	true	Screenshots to KV store on failure
maxRetries	integer	2	Retries per (platform, question)
askTimeoutMs	integer	90000	Max wait per answer
useApifyProxy	boolean	true	Route through Apify Proxy
cookies	array	[]	Session cookies for authenticated platforms

Input example

{ "questions": ["best CRM for startups"] }

Output

Dataset record (one per question × domain)

{ "question": "best CRM for startups", "domain": "g2.com", "count": 12, "category": "review", "platforms": ["ChatGPT", "Perplexity"], "urls": ["https://www.g2.com/..."] }

Key-Value Store artifacts

• SUMMARY — ranked top-lists: topDomains, topReviewSites, topBlogs, topCommunities, topForums, topSocialMedia, topGovernmentDomains, topDocumentationSites, topNews
• REPORT — final ranked markdown report
• EXPORT_CSV — flat CSV of all records

Domain categorization

Domains are normalized to their registrable form (www.G2.com/x → g2.com, apple.co.uk handled) and classified into: review, blog, community, forum, social, government, documentation, news, ecommerce, other.

Rate limits & known limitations

• Perplexity is citation-first and most reliable; ChatGPT/Claude/Gemini often need cookies.
• Citation coverage depends on whether the engine surfaces links for a given answer.
• Selectors live in src/shared/adapters/* and can be updated without touching the pipeline.

Deployment

npm install && npm run build && npm test && npm run lint
npx playwright install chromium   # local only; Docker image already has it
apify push

Docker base image: apify/actor-node-playwright-chrome:22.

Testing guide

npm test runs citation/parser (fixture-based), retry, validation, statistics, adapter (mock page), engine integration, and analysis tests. Fixtures live in tests/fixtures/.

Troubleshooting

Symptom	Cause	Fix
Few/zero citations	Engine returned no links / auth wall	Provide cookies, prefer Perplexity
error: captcha	Bot check	Enable residential proxy, retry

Version history

• 1.0.0 — Initial production release.