How Citations.io works
Everything you need to set up your brand, read your results and act on them - written in plain English first, with the precise detail underneath for search and analytics teams.
Quick start
Citations.io measures how often AI assistants name, recommend and cite your brand when people ask questions in your category. Getting your first result takes about ten minutes.
- 1Create your accountStart a 5-day trial for $5 or sign in with Google. One account can hold multiple brands on higher plans.
- 2Add your brand and domainYour domain is used to identify your brand in answers and to detect when an engine cites one of your own pages.
- 3Confirm your competitorsWe suggest competitors from your domain and category. Review them - accurate competitors make share-of-voice meaningful. You can change them at any time.
- 4Review your promptsWe seed a starting prompt set from your brand and market. Edit, delete or add your own before scanning.
- 5Run your first scanScanning runs in the background across ChatGPT, Gemini, Perplexity and Claude. You can close the tab; results appear on your dashboard when it finishes.
Prompts
A prompt is a question a real buyer might type into an AI assistant - for example "best project management tool for agencies". Prompts are the unit everything else is measured against: your score, your competitors' share of voice and your recommendations all come from the answers your prompts produce.
Your plan sets how many prompts you can track: Starter includes 25, Growth 75 and Pro 150, each with a matching allowance of benchmark prompts. Add prompts from the Prompts screen, or promote a suggested prompt we've spotted in your category.
Prompts are locked into a snapshot at the start of each monthly cycle. That snapshot is what gets scanned, so month-on-month comparisons stay like-for-like. Changes you make mid-cycle are promoted into the next cycle rather than altering the one in progress.
Observations
An observation is a single AI answer we captured for one of your prompts on one engine, stored verbatim with everything we detected in it: whether you were mentioned, which competitors appeared, the sources cited, the position you appeared in and the tone used about you.
Each prompt is sampled repeatedly across ChatGPT, Gemini, Perplexity and Claude, because AI answers are non-deterministic - asking the same question twice can produce different brands. Repeat sampling is what turns a one-off answer into a reliable measurement.
On the Observations screen your prompts are listed as a table with visibility per prompt. Expand a row to read every sampled answer behind it, and use the filters at the top to narrow by engine, by whether your brand was mentioned, or by free-text search.
Scores and metrics
What it means
How often your brand shows up in AI answers for your tracked prompts, weighted by how prominently it appears.
What to do
Track the direction, not the decimal. A rising score means AI assistants are naming you more often in the questions you care about.
What it means
The same measurement split by ChatGPT, Gemini, Perplexity and Claude, because each engine draws on different sources.
What to do
Find your weakest engine and look at what it cites - that tells you which sources to go and win.
What it means
The brands named in an answer, including you and your competitors.
What to do
Where a competitor is mentioned and you are not, that prompt is a concrete content or PR target.
What it means
The websites an engine linked to or drew on when writing the answer.
What to do
Recurring third-party domains are the pages worth being featured on - outreach here moves visibility faster than on-site work alone.
What it means
Whether an answer describes your brand positively, neutrally or negatively.
What to do
Negative or inaccurate framing usually traces back to one bad source; find it in the citations and correct it.
What it means
How much sampling sits behind a number. More observations, higher confidence.
What to do
Treat low-confidence movement as noise; wait for the next cycle before acting on it.
Competitors and benchmarks
Share of voice compares how often you are mentioned against your named competitors across the same prompts and the same observations. Because everyone is measured on identical questions, it is a fair market-relative read rather than an absolute one.
Benchmark prompts are category-level questions that aren't specific to your brand - they show how your market behaves overall and where there is space to win. They use a separate allowance, so adding benchmarks never eats into your own prompt limit.
Implementation Packs
Each cycle we turn your results into a prioritised set of actions: the prompts you're losing, the sources driving competitor mentions, and the on-site pages worth writing or rewriting. Each recommendation names the evidence it came from, so you can hand it straight to a content, PR or dev owner.
Work through the queue top-down - items are ordered by expected impact on visibility, not by effort.
Reports and exports
Reports summarise a cycle: headline visibility, engine breakdown, competitor share of voice, notable citations and what changed since last month. Trials include one full scan and dashboard access on your own brand; paid plans add full reports and data exports, with export breadth increasing on Growth and Pro.
Billing and plans
Trials run for five days for $5 and cover one brand with a limited prompt set so you can see real results before committing. When a trial converts, your prompt allowance jumps immediately - most brands under-use it in the first month, so revisit your prompt list the day you upgrade.
Plans are billed monthly: Starter (25 prompts), Growth (75) and Pro (150), each with a matching benchmark allowance. Upgrades apply straight away; the extra allowance is available in the current cycle. Manage plans, invoices and cancellation from Settings → Billing.
Troubleshooting & FAQ
My scan says pending - how long does it take?+
I have no observations yet - what's wrong?+
The competitors listed aren't my competitors.+
One engine shows no results.+
How do I get help?+
Start a 5-day trial for $5 and run your first scan today.
Start your trial
