1. Our Mission
GetCited.me is a Generative Engine Optimization (GEO) platform. We help brands understand how AI assistants mention, cite, and recommend them — then turn those gaps into concrete work: visibility tracking across answer engines, GEO audits, citation-ready content drafts, and community reply drafts.
This page explains how we interact with AI platforms, what data we use, and what we deliberately refuse to do. It complements our Privacy Policy and Terms of Service.
2. How We Interact with AI Platforms
For visibility tracking, we query AI platforms through authenticated, paid, commercial APIs — not by scraping chat websites:
- Google Gemini API — Gemini answers, requested directly from Google's own Gemini API
- OpenRouter — ChatGPT (OpenAI GPT), Claude (Anthropic), and Perplexity (Sonar) answers by default. OpenRouter is a commercial API router: it forwards each request to a provider serving that model, with web search turned on as it is on the providers' own endpoints. We can switch any of these three engines back to its provider's own API (OpenAI, Anthropic, or Perplexity), in which case the same request goes there instead
For Google AI Overviews visibility signals, we use DataForSEO’s commercial SERP API — a paid third-party data provider — not Google’s consumer UI and not unofficial scrapers. Google AI Mode is not a tracked surface in GetCited.me today.
Some research steps — the brand and competitor analysis behind your brand profile — also call the Gemini and Perplexity APIs with your brand name, website address, industry, brand description, and the paths of your site's pages.
For content briefs, full articles, GEO audit analysis, topic discovery, community draft writing, the in-app assistant, translation of articles you send to Prismic (currently a Google Gemini model), and speech-to-text for voice dictation, we route model calls through OpenRouter under our paid account. Those jobs receive only the working context needed for the task (for example public page text you asked us to analyze, brand description, cited source URLs, the article being translated, your messages to the assistant, or your dictation audio) — not your password, billing credentials, or private account settings.
All of the above usage is subject to each provider’s terms of service. We maintain authenticated access and operate within documented rate limits and quotas.
3. No AI-Platform Scraping or ToS Circumvention
We want to be explicit about what we do not do when measuring AI visibility:
- We do not scrape ChatGPT, Claude, Gemini, Perplexity, or similar chat websites or user interfaces to collect answers
- We do not reverse-engineer proprietary AI client protocols to bypass official access
- We do not use unofficial or undocumented endpoints for tracking runs
- We do not circumvent provider rate limits or abuse API quotas
Tracked questions are sent as ordinary API requests that contain only the text of the question — no brand context, account data, or personal data. The responses we store are the same class of publicly generated answers a user would receive for the same question.
4. Public Data Only
The AI answers we analyze are publicly generated model outputs for the questions you track (created by you, or suggested by our product and approved by you). We do not access private chats, enterprise workspaces, or restricted corpora belonging to other customers of those platforms.
We fetch only publicly accessible pages of your website — the same class of content available to search engines and ordinary visitors — and we do not enter authenticated areas, paywalled sections, or admin panels. Those pages are fetched:
- When you add the website or ask for a scan, an audit, a brand-profile refresh, or a content job
- In the background, to keep the stored list of your public pages current: it is refreshed when it is more than two weeks old, including during a monthly pass over every website that adds internal links to the articles you have published
- When an AI answer cites one of your pages: we open cited public pages, yours included, to resolve their address and title
Scheduled tracking runs on your plan's cadence (weekly on the free plan, up to daily on paid plans) and only asks AI platforms your tracked questions; it does not crawl your site. If you connect the WordPress plugin, it reports an inventory of your published pages from inside WordPress shortly after pairing, then once a day and whenever an administrator asks for a sync — exactly what it sends is listed in our Privacy Policy.
5. Community Threads & Reply Drafts
GetCited.me can discover public discussions (for example on Reddit, Stack Overflow, or Hacker News) that are relevant to your brand, and generate organic reply drafts for threads you choose.
Important boundaries:
- Discovery and drafting use public thread metadata and content.
- Where a platform offers an official API, we prefer it. For Reddit draft quality, when APIs alone are insufficient, we may load the public thread page in a controlled browser session only to gather context for a draft you requested — not to automate posting, not to mass-message users, and not to interact with AI chat UIs.
- If you add the public Reddit profile you reply from, we read that profile's public activity once a night to estimate whether your replies are likely to survive moderation. We never ask for your Reddit password.
- We generate drafts for your review. Publishing or posting remains under your control (copy/export or platform handoff). We do not silently post comments or messages as you.
- Drafts are meant to be thread-aware and useful — not spammy brand pitches. Low-quality or inappropriate threads can be filtered out of recommendations.
6. No Unsolicited Marketing or Outreach
GetCited.me is a monitoring, audit, and content-assistance product. We are not an outreach agency:
- We do not send emails, DMs, or any outreach to third parties on your behalf — the only email we send to someone else at your request is the invitation to a colleague you add to your team
- We do not run advertising campaigns or promotional blasting as part of the Service
- We do not contact competitors, publishers, or community members about you without your own action
- We do not submit generated content into AI platforms to influence their rankings or citations
Any action based on our recommendations — publishing an article, posting a community reply, changing your site — is entirely at your discretion.
7. Content Generation & Publishing
We generate content briefs, article drafts, recommendations, and related artifacts using commercial AI APIs (see section 2). Drafts are created for your use and review, and you retain control over whether and how to edit or publish anything we generate.
- WordPress — with the GetCited.me plugin connected, Create draft in WordPress has the plugin create a draft post on your site. Scheduling and publishing happen in WordPress, by you; the plugin then reports the post's status back to us, never its content.
- Prismic — sending an article creates draft documents in your repository's Migration release through Prismic's Write API, and you publish them in Prismic. If you choose extra locales, the article is first machine-translated by a Google Gemini model through OpenRouter.
- Other destinations (Dev.to, Medium, Substack, your own site, plain Markdown) — the article is copied to your clipboard and, where there is one, the destination's editor opens in a new tab. You can then mark the draft as published in GetCited.me.
- We never publish to your CMS or communities on our own: each destination above needs an explicit action by you, in the product or in your CMS.
- Generated drafts belong to you under our Terms of Service; providers process requests under their API terms.
8. Images & Human Creativity
Article posters and similar imagery in the product are either uploaded by you or sourced from Unsplash via their official API when that feature is enabled. We show photographer attribution where Unsplash requires it.
We do not generate AI images for those posters. Real photography from real creators is intentional: supporting human work is part of how we want to ship AI-assisted content tools. If we add generative imagery later, it will be disclosed here and designed to complement — not silently replace — human creative attribution.
9. Data Storage & Security
Account and product data are stored in our application database with encryption in transit (TLS). Passwords are stored only as salted bcrypt hashes. Session tokens expire and rotate. The access tokens you issue for our MCP server or the WordPress plugin are stored only as one-way hashes, work only within the scopes and websites you choose, expire within 365 days, and can be revoked at any time. Integration secrets — your Search Console authorization and Prismic tokens — are encrypted at rest (AES-256-GCM). Production access is limited to authorized personnel.
When you delete a website or your account, we revoke the access tokens bound to it (including the WordPress plugin's credential), ask Google to revoke the Search Console authorization, and delete the stored integration secrets before the data itself is removed. Prismic offers no way for us to revoke its write token, so delete it in Prismic if you no longer need it.
We do not claim perfection — no online system is 100% secure — but we apply industry-standard controls appropriate to a SaaS analytics product. Full details on what we collect, retention, and your rights (including GDPR where applicable) are in our Privacy Policy.
Model training. Requests we route through OpenRouter go only to endpoints whose providers' policy is not to train on request data: our OpenRouter account excludes every endpoint, paid or free, that may train on request data or publish it. This is an account setting rather than something our code enforces on each request, and we do not require Zero Data Retention, so a provider may still keep request data for a limited time under its own policy — for example to monitor abuse. Requests we send directly to a provider's own API (Google's Gemini API, Perplexity, or an engine switched back to its own provider) are governed by that provider's commercial API terms. We do not use your data to train AI models of our own.
10. Contact
Questions about AI ethics, platform access, community drafts, or data practices: support@getcited.me.