LLMrefs review (2026): one €79 plan, unlimited domains, and a free scan before the paywall

LLMrefs sells AI-search rank tracking on a single plan: All in One at €79/month with a 7-day free trial, 500 tracked prompts, weekly refreshes, and — unusually for this category — unlimited projects and unlimited team seats. The standout move is what happens before the paywall: it runs a real sample scan of your prompts live, with a progress bar, and shows you a ranked share-of-voice preview before the paywall.

What LLMrefs is

LLMrefs tracks how often AI engines name your brand when answering category prompts. The two core metrics are Share of Voice (mention rate across all AI responses) and Position (average placement of the first mention). It claims coverage of "ChatGPT, Google AI Overviews, Gemini, Perplexity, Microsoft Copilot, and more" — the onboarding shows eight engine icons, but only those five are named in text. Method transparency is a genuine plus: the tool says it queries engines with "fresh browser sessions to capture the complete response text," and you can open the original AI answer behind every metric.

The onboarding, minute by minute

Setup asks you to pick Brand or Agency mode, enter a brand URL, name, location and language, then suggests five topics for your niche and generates 10 prompts per topic — 50 prompts to start, all editable. Then it runs the scan in real time: "We're querying AI search engines in real-time. This usually takes a few minutes," with a live percentage.

The preview we got back was a ranked leaderboard: our test brand placed 5th in the United States with an 18% share of voice, in a field where the tool found 174 competing brands. Rows below rank 6 are blurred behind a "click to unlock" overlay — a fair trade for a scan that cost us nothing. One detail worth knowing: the ranking is not sorted by share of voice alone. A brand with 28% SoV sat at #3 behind one with 25%, because average position weighs in.

Pricing and limits

The plan card lists: 500 prompts tracked for brand mentions, sources and fan-out queries; weekly auto-refreshed reports; geo-targeting in 50+ countries and 20+ languages; unlimited projects under one subscription; CSV export and API access; all AI engines with no additional fees; citation tracking; unlimited team members; monthly AI prompt-volume estimates; priority support. It also bundles side tools: an AI crawlability checker, a Reddit threads finder, and an llms.txt generator.

For agencies, "unlimited projects, one subscription for all your domains" is the headline. Most competitors meter by brand or workspace; here the €79 covers the whole client roster — on paper. We didn't load-test that claim.

What we liked

• Data before money. The free preview scan is real engine output, not a demo. Nobody else we tested shows you your own numbers before the paywall.
• Transparent method. Original AI responses are viewable behind every metric — you can audit the tool's claims.
• Prompt-volume estimates (beta). It showed, e.g., "generative engine optimization" at roughly 200 monthly AI conversations in the US — directional, but a useful sanity check on what's worth tracking.
• They practice what they sell. llmrefs.com serves an exemplary llms.txt — 180 links across 6 sections, every one with a description. It passes our own 5-signal check with a perfect score.

What bugged us

• One auto-generated prompt read "What are the top AI search monitoring platforms in 2024?" — in a test run mid-2026. Stale template prompts quietly skew tracking.
• Weekly refresh only. If you ship a fix on Monday, you wait for the next cycle to see movement.
• The full engine list is unclear: five named, eight icons, "and more."
• "Limited time only" pricing with no visible annual option or agency tier makes budgeting a guess beyond the first month.

What we couldn't test

The 7-day trial gates the full dashboard, so we didn't see prompt-level drill-downs, citation reports over time, or the CSV/API output. Whether the trial requires a card wasn't visible either. If you've run LLMrefs in production, we'd genuinely like to hear how the weekly data holds up.

Verdict

The strongest value story of the four tools we tested this week: one honest price, no per-domain metering, and the confidence to show you real data before charging you. The trade-offs are cadence (weekly) and depth we couldn't verify past the paywall.

If €79/month is more than your problem justifies, start free: run the 5-signal check on your domain, then track movement with Monitor from $20/month . Same discipline, smaller bill.

Also in this series: Otterly review , Profound pricing teardown , our Scrunch test that got blocked .