There are two very different questions behind "AI citation analysis", and tools answer one or the other. The first: are my pages cited as sources in AI answers. The second: which sources do the engines lean on when they answer my market's questions, whoever they belong to. The first is a scoreboard. The second is a map, and it is the one that tells you where to go next.
This page compares the tools that do this work, on facts each vendor publishes, read on 2026-08-18. Epovest is one of them, and the table is ordered by published entry price. The same platforms are compared on AI visibility as a whole in our general comparison, on the answer engines they cover in our guide to AEO tools, and on what a year of measurement costs in our guide to GEO tools.
What a citation analysis has to produce
The source behind the answer, not just the count. An engine names a brand because something it read said so. The useful output is that something: the domain, the page, the date it was seen.
A per-engine reading. The sources differ sharply from one engine to the next; a list merged across engines hides exactly what you need to act on. Epovest scores each source per engine on a 0 to 100 scale (AI Authority), because a domain that carries Perplexity may be invisible on ChatGPT. One domain, measured four ways: reddit.com is Gemini's most cited domain and three citations out of 1,945 on ChatGPT.
A path back to the raw text. A citation you cannot open is a claim. Where the answers are archived with their sources, any figure can be reopened months later. The tools that state they archive them, and at what price, are in our guide to LLM tracking tools.
Something to do with it. A map is worth what it changes: which pages to publish, which third-party pages to get onto, which of your own surfaces to align first.
The comparison
| Tool | Entry price (published) | Engines named | Raw answers kept, per the vendor's pages |
|---|---|---|---|
| Siftly | $0/month, then $79/month | 9 | Response counts per plan |
| Keyword.com | $7.83/month billed annually | 9 | Not specified |
| AEO Vision | $9/month, 1 engine at that tier | 5 | Not specified |
| Epovest | $10 minimum top-up, then $0.10 per prompt, per engine, per survey | 6, all through the engines' official APIs | Every answer archived with its cited sources, and every source ranked per engine in the Atlas |
| Rankscale | From $20/month | 10 | Not specified |
| Ayzeo | $39/month, engines sold separately | 9 | Not specified |
| RankPrompt | $470/year billed annually | 7 | Not specified |
| LLM Pulse | EUR 49/month | 11 | Not specified |
| AmICited | EUR 50/month | 7 | Full answer archive listed |
| AI Clicks (aiclicks.io) | $59/month | 10 | Not specified |
| Omnia (useomnia.com) | EUR 63/month | 7 | Not specified |
| Mentionable | EUR 79/month | 7 | Not specified |
| Dageno | $79/month | 8 | Not specified |
| Mint (getmint.ai) | $80/month | 4 | Not specified |
| Rankfender | $89/month | 6 | Not specified |
| Profound | $99/month billed yearly | 9 | Not specified |
| Sight (trysight.ai) | $99/month | 7 | Not specified |
| Scrunch | $250/month | 8 | Not specified |
| Athena | $295/month | 8 | Not specified |
| Cognizo | From $499/month | 8 | Not specified |
| Evertune | $800/month | 11 stated | Not specified |
| Peec AI | Not published on the page read | 8 | Not specified |
| Geneo | Not published on the page read | 4 | Not specified |
"Not specified" describes the vendor's public pages on the read date, and nothing about what the product does.
What the map looks like once you have it
A worked example from our own measurement, published so the shape is clear rather than abstract. On the travel questions we measure, reddit.com scores 100 on Gemini, Perplexity and Grok, and 28 on ChatGPT; lonelyplanet.com scores 100 on ChatGPT and 22 on Gemini. Same market, same week, two engines that barely agree on what to read.
That is the whole argument for a per-engine reading: a brand that decides where to appear from a merged list optimises for an average that no engine actually uses. The full study behind those figures is here.
What each tool is, per its own pages
Epovest archives every answer with its cited sources and ranks those sources per engine in the Atlas, on a 0 to 100 AI Authority scale computed across every measurement it runs, which is what makes two accounts' figures comparable. Credits are debited only by the surveys that run, at $0.10 per prompt, per engine, per survey, with no expiry. The REST API and MCP server ship with every account, so your own assistant reads the map and works from it.
AmICited lists a full answer archive among its features, from EUR 50 per month across seven named engines.
Siftly publishes a free tier and counts in tracked responses per month. Keyword.com and AEO Vision hold the low entry prices, with the engine restrictions noted in the table.
Rankscale, Ayzeo, RankPrompt, LLM Pulse, AI Clicks, Omnia and Mentionable publish entry prices from $20 to EUR 79 per month.
Dageno, Mint, Rankfender, Profound and Sight publish from $79 to $99 per month. Scrunch, Athena, Cognizo and Evertune publish enterprise-facing prices from $250 to $800.
Peec AI and Geneo did not publish an amount on the pages read.
Method
Selection: the platforms named or cited by ChatGPT, Claude, Gemini, Perplexity, Mistral and Grok on this market's buying questions, measured on our own panels, weekly, in English and French, every answer archived. No tool was added or removed by hand.
Facts: read on each vendor's public pricing and product pages on 2026-08-18. Base prices only; promotional discounts are ignored. The AI Authority figures quoted above come from our own published measurement, with the study linked.
Order: published entry price, ascending, at its nominal amount in the currency the vendor publishes. The page contains no score and no ranking.
Figures may be reused with attribution.