Hikoo or Qwairy: the 2026 numbers compared

Hikoo or Qwairy: the one-minute answer

Qwairy suits teams that want to cover the entire French AI landscape, Mistral and Claude included, without trading engines off against each other. Hikoo suits organisations that track ChatGPT and Gemini closely, publish content on a regular cadence, and prefer a model with no credits to work out.

A useful disclosure before going further: I co-founded Cockpyt AI, a platform that competes with both. So you are reading a comparison written by a market participant. I pulled both pricing grids from the source on 13 August 2026, and every calculation in this article can be reproduced from their public figures. Cockpyt appears neither in the table nor in the verdict.

Hikoo vs Qwairy: the numbers side by side

Both vendors are French, host in Europe and publish their prices. The comparison therefore rests on what actually differs: engines included, volume, and the consumption model.

On the headline price line, Hikoo is consistently €10 to €60 cheaper than the equivalent Qwairy tier. On every line below it, the gap reverses.

How many complete scans does your subscription really allow?

Neither platform supports daily tracking at full prompt load, despite what their pricing pages advertise. The calculation runs on their own published figures.

A complete pass means your prompts queried once on each active engine. On Hikoo that consumes AI answers. On Qwairy it consumes credits, the documented rule being one credit per monitored answer per model. Divide the monthly quota by the cost of one pass and you get your real refresh rate.

On Hikoo, the refresh rate falls as the price rises

• Lite: 30 prompts across 2 engines, so 60 answers per pass. The 1,500 quota allows 25 passes per month.
• Growth: 80 prompts across 4 engines, so 320 answers per pass. The 5,000 quota allows 15 passes.
• Scale: 200 prompts across 7 engines, so 1,400 answers per pass. The 15,000 quota allows 10 passes.

The €389 tier therefore delivers fewer complete updates than the €69 one. The logic makes sense, since each pass costs more as the scope widens, but the result contradicts the daily update label shown on all four plans. Genuinely daily tracking on Scale would require 42,000 answers per month, close to three times the included quota.

On Qwairy, the budget is 13 credits per prompt per month

The Qwairy grid follows a constant rule: 100 prompts for 1,300 credits, 300 for 3,900, 800 for 10,400. Every tracked prompt gets exactly 13 monthly credits. Across the 10 included engines, one complete pass consumes 10 of them, leaving 1.3 complete passes per month on all three plans.

The three selling points on the pricing page, 10 engines, 100 prompts and daily frequency, are therefore each true and mutually incompatible. You pick two. Weekly tracking on 100 prompts means dropping to 3 engines. Tracking across 10 engines means dropping to roughly 30 prompts. The pricing page does not lie, it simply leaves the trade-off to you.

Two points in Qwairy’s favour. Unused credits roll over to the following month, which smooths out peaks. And engine selection happens analysis by analysis, so nothing forces you to query all 10 every time. The constraint sits on the overall budget, not on flexibility.

The limits of this calculation

These figures rest on two reading assumptions. On Hikoo, one AI answer corresponds to one prompt queried on one engine. On Qwairy, the current FAQ states that consumption varies by engine and operation, which makes my estimate a ceiling: the real pass count can be lower, never higher. A content brief drawn from the same counter reduces the measurement budget accordingly.

In my view, this calculation is the first one to run before comparing two GEO subscriptions, and nobody publishes it. The advertised prompt count is a configuration ceiling, not a measurement volume. Two platforms both advertising 100 prompts and daily updates can deliver refresh rates twenty times apart. That number determines how fine your curve will be, and that is what you should be negotiating.

Why frequency matters less than the pricing grids suggest

The inherent variability of these models limits what closely spaced tracking can actually detect. Ask a model the same question twice and the answers differ, including at zero temperature with a fixed seed. A 2025 arXiv paper generated 1,000 responses to an identical prompt with identical seeds on GPT-4o-mini and observed that “approximately 95% of nodes have exactly one child”, the remaining 5% being generation forks. The work covered model weight exfiltration detection rather than brand citation measurement, but the underlying mechanism is the same.

Part of what you measure daily is therefore sampling noise, not movement in your visibility. That visibility depends on published content, citations earned elsewhere and index updates, all of which move on a weekly scale. A tighter rhythm at constant prompt volume buys variance, not information.

The trade-off looks different once you accept that. On a fixed monthly budget, favour prompt volume over frequency, because 100 prompts measured once produce a steadier observation base than 20 prompts measured five times. On that ground, Qwairy’s generosity on prompts outweighs its refresh rate.

Where Hikoo takes the lead

Three situations point to Hikoo without hesitation.

Tracking ChatGPT and Gemini at the best price. Those two engines account for 86.3% of referral traffic sent by AI to websites, according to the measurement SE Ranking published in June 2026 across a panel of more than 100,000 sites. Hikoo Lite covers both for €55 on annual billing. The equivalent Qwairy plan costs €65 and bills you for 8 engines you will not use.

Content production on a regular cadence. Hikoo keeps the counters separate: 30 article suggestions per month on Growth do not eat into your measurement budget. On Qwairy, a Content Studio brief draws from the same credits as monitoring. On a Starter plan, producing ten briefs can absorb a significant share of the monthly measurement budget. If your primary use is editorial, Hikoo’s separated quotas avoid a permanent trade-off.

Extended testing before deciding. Hikoo’s Free plan is permanent: 10 prompts, 50 answers per month on ChatGPT, 100 audited URLs, 2 competitors, 1 article suggestion. Lite and Growth add 7 days of full access with no credit card. Qwairy offers 120 free credits, which amounts to a single pass across a dozen prompts. To evaluate a tool over several weeks before committing budget, Hikoo offers the widest front door on the French market.

On top of that sits the free tool suite the vendor publishes, including an llms.txt generator and an AI crawler access checker, both usable without an account.

Where Qwairy takes the lead

French market coverage. Mistral and Le Chat appear on Hikoo only on the Scale plan at €389. On Qwairy, Mistral is included from €65. For a French brand whose audience uses Le Chat, the billing gap to reach the same engine reaches €244 per month. Claude, DeepSeek, AI Overviews and AI Mode follow the same pattern.

Prompt volume. 100 prompts on Starter against 30 on Lite. A single-market business with three product lines and two personas passes 30 prompts without effort. That ceiling forces a move to Hikoo Growth at €139 annual, against €65 at Qwairy for a wider scope.

Multi-brand management. Qwairy publishes a workspace count per plan, 1 then 5 then 20, with unlimited team members. Hikoo does not publish this on its pricing page and points to a quote-based agency offer. For a consultant or an agency, that opacity makes budgeting harder.

Transparency on access mode. Qwairy states engine by engine whether measurement runs through the interface or the official API: interface only for Copilot, AI Overviews and AI Mode, API only for Claude, Mistral and DeepSeek, both for ChatGPT, Perplexity, Gemini and Grok. That distinction changes how reproducible the measurement is and how it sits with each vendor’s terms of use. Hikoo does not document this publicly.

In my view, Hikoo’s per-engine tiering is its main strategic problem, more than its price. Reserving Mistral for the top tier in a market where Le Chat carries weight among French-speaking users forces a French customer to pay for Grok and Claude in order to reach a local engine. Qwairy went the other way, including everything everywhere and charging for volume rather than scope. The second approach strikes me as better aligned with how the market actually consumes these tools.

Which tool fits your profile?

• Micro-business or independent tracking one brand on ChatGPT and Gemini. Hikoo Lite, €55 annual. The scope matches the need and the Free plan lets you validate before paying.
• French brand whose audience uses Le Chat or Claude. Qwairy Starter, €65 annual. It is the only one of the two to include Mistral under €389.
• Editorial team publishing weekly. Hikoo Growth. Separating the content quota from the measurement quota avoids choosing between writing and measuring.
• Consultant or agency handling 3 to 20 brands. Qwairy Growth or Business, for the published workspaces, the REST API and the Looker Studio connector.
• Organisation requiring genuinely daily tracking. Neither, on the published plans. Negotiate an Enterprise quote specifying your monthly pass count, not your prompt count.

The third option if neither fits

Cockpyt AI, which I co-founded with Laurent Séjourné, starts from a different trade-off: three engines only, ChatGPT, Gemini and Perplexity, but a fixed weekly rhythm and a monthly deliverable. Those three engines carry 93.6% of the AI referral traffic SE Ranking measured. Mistral, Claude and AI Overviews are not tracked: if covering the whole French AI landscape is your priority, Qwairy remains the right pick of the two compared here.

The difference sits in the deliverable. Every month and per project, a GEO audit combines the technical analysis of your site from an AI perspective, a study of 10 questions tested across several phrasings, an analysis of AI-sourced traffic, a prioritised action plan and a 33-point checklist. All eight modules are included on all three plans, with only the volumes changing: Solo at €29 per month annual for 1 project and 15 questions, Consultant at €79 for 3 projects and 45 questions, Agency at €139 for 5 projects and 75 questions. The trial runs 14 days with no credit card.

The refresh rate, calculated the same way as above, is 4.33 guaranteed monthly passes across all questions and all three engines, with no credit budget to arbitrate. That stability is paid for with a narrower engine scope, a trade-off I state openly.

Frequently asked questions about Hikoo and Qwairy

Does Qwairy really track 10 engines on its entry plan?

Yes, with no add-on and no surcharge. ChatGPT, Perplexity, Copilot, Gemini, AI Overviews, Grok, Claude, Mistral, AI Mode and DeepSeek are available from the Starter plan. The constraint is not access but the credit budget: querying 100 prompts across all 10 engines consumes 1,000 of the 1,300 monthly credits included.

Can you track Mistral on Hikoo without moving to Scale?

Not on the published plans. Mistral appears only among the 7 platforms of the Scale plan at €389 per month. The vendor offers additional model add-ons on any subscription, including AI Overviews and AI Mode, but their price is not published and requires contacting sales.

Is Hikoo’s free plan enough to get started?

For a first snapshot, yes. It covers 10 prompts and 50 answers per month on ChatGPT alone, roughly 5 monthly passes, with 100 audited URLs and 2 competitors. For tracking a trend across several engines, the volume becomes limiting within weeks.

How do you convert Qwairy credits into a prompt count?

Multiply your prompt count by the number of active engines to get the cost of one pass, then divide the monthly quota by that result. Each plan allocates 13 credits per prompt per month. Real consumption varies by engine and operation, so treat the result as a maximum.

Can you migrate your history from Hikoo to Qwairy?

No. Visibility scores are proprietary and calculated with different weightings, so a score of 62 at one vendor does not equal 62 at the other. Export your data before cancelling and run both tools in parallel for a month to establish a conversion factor on your own prompt set.

Which of the two suits agencies better?

Qwairy on the published plans, thanks to the 5 workspaces on Growth, the 20 on Business, unlimited members and the Looker Studio connector. Hikoo has an agency offer with preferential pricing and a white-label kit, but it goes through a quote, which prevents any numerical comparison at this stage.

Do you need daily tracking of your AI visibility?

Rarely. Model answers vary from one query to the next even at identical settings, while your real visibility moves at the pace of published content and earned citations. On a constant budget, increasing the number of prompts measured stabilises the reading more than multiplying passes over a smaller sample.

Sources

• Hikoo, official pricing page, accessed 13 August 2026. tryhikoo.com/fr/tarifs
• Qwairy, official pricing page and plan comparison table, accessed 13 August 2026. qwairy.co/pricing
• SE Ranking, Analysis of Top AI Search Engines: Who Is Catching Up to ChatGPT?, June 2026. seranking.com/blog/ai-traffic-research-study
• Verifying LLM Inference to Detect Model Weight Exfiltration, arXiv:2511.02620, 2025. Empirical characterisation of non-determinism across 1,000 fixed-seed generations. arxiv.org/pdf/2511.02620
• Dissecting Non-Determinism in Large Language Models, ICLR Blogposts, 2026. iclr-blogposts.github.io