Nightwatch Review 2026: Thirty Answers a Month, Whatever You Pay

Nightwatch is a rank tracker that added AI visibility and charges nothing extra for it, and it is the second tool we have found whose published numbers let you derive its sampling. Prompts and total AI responses sit in the same column on every tier. Two divisions and you know what you are buying.

What they give is a constant: thirty answers per prompt per month on every plan, from the €79 one to the €399 one. More money buys more prompts, never more depth.

Disclosure: EchoWi competes with Nightwatch. Every price, keyword count, prompt count, response count and engine name below was read from Nightwatch’s own pricing page on 7 August 2026, and the arithmetic is shown so you can repeat it.

The short version

1. From €79 a month, with a 14-day trial and no card required.
2. AI tracking is included on every tier at no extra cost. It is not an add-on.
3. Claude on the entry plan, which a third of this market charges more for.
4. Thirty AI responses per prompt per month, on every tier. Derived, not stated.
5. Unlimited seats on every plan, which is genuinely unusual.

What it costs, verified

Annual billing takes 20% off. Every tier names the same AI engines, ChatGPT, Claude, Gemini and Perplexity, plus AI Overviews on Google, YouTube and Bing. Seats are unlimited everywhere, and the page states no setup fees, no long-term contract and cancellation at any time.

The constant hiding in the table

Divide responses by prompts on each row:

Thirty, three times. That is not a coincidence, it is a design decision: every plan collects thirty answers per tracked prompt per month, and the ladder sells breadth. Going from €79 to €399 multiplies your prompt count by ten and leaves the depth of each one exactly where it was.

Almost nobody in this market lets you see that. It is the second time we have been able to derive a sampling rate from a public page, after Siftly, and the two vendors made opposite choices: Siftly’s depth rises with the tier, from one run a day to three, while Nightwatch’s is flat.

What thirty answers a month actually supports

Thirty per prompt, spread across the engines. The page names four models plus AI Overviews on three surfaces, so the divisor is somewhere between four and seven depending on how you count.

• At four engines: 7.5 answers per engine per month, roughly weekly.
• At seven: 4.3 per engine per month, roughly fortnightly.

Now the uncertainty. For a mention rate near 30%, a 95% confidence interval at n=7.5 is about plus or minus 33 points. At n=4.3 it is about plus or minus 44. In other words, a single prompt on a single engine tells you almost nothing month to month, and the arithmetic behind that is the same one we apply to everyone.

That is not a Nightwatch failing. It is what thirty answers buys, and Nightwatch is one of the only vendors that lets you work it out before paying. The practical consequence is the same advice we give everywhere: read the aggregate across your whole prompt set, not the per-prompt rate. Five hundred prompts at thirty answers each is 15,000 observations, and that number is usable even though none of its components is.

What an answer costs here

Set against the rest of the comparison, that is mid-table: cheaper than Rankscale at $0.021 only on the Agency tier, well under Athena HQ at $0.082, and more expensive than Siftly at $0.0055 on its top tier.

But the comparison is unfair in Nightwatch’s favour and it is worth saying why. The €79 also buys 500 keyword rankings tracked daily, five websites, site audits and an SEO agent with 100 credits. If you need rank tracking anyway, the AI half arrives at no marginal cost and the per-answer figure is meaningless. If you do not, you are paying rank-tracker prices for 50 prompts.

Claude on the entry plan

Worth its own line, because it is rare.

Two tools in this comparison put Claude behind an Enterprise tier with no published price, Cognizo and Siftly, and two more charge for it on a middle tier: Surfeo at €99 and Knowatoa at $199. Writesonic reserves seven of its ten platforms for an unpriced Enterprise tier without naming which. Nightwatch names Claude on its €79 plan, alongside ChatGPT, Gemini and Perplexity, and the engine list does not change as you climb.

Promptmonitor, Rank Prompt, Waikay, Alhena and Athena HQ make the same choice, the last of them putting Claude on its free tier. Counting the whole catalogue rather than an impression, six vendors put Claude on the plan you can actually buy and four charge more for it, which is the opposite of what we assumed.

Rank Prompt goes further and inverts the usual line: Claude sits inside its base credit while Gemini costs extra, which is the only case of that we have found.

The third shape this market comes in

Cataloguing this market has made a split visible that no vendor describes, and Nightwatch sits squarely in one half of it.

Tools built for AI visibility from nothing. Profound, Peec AI, Otterly, Scrunch, Rankscale, Superlines, Athena HQ, Siftly, Cognizo, Dageno, Rank Prompt, Evertune, Brandlight, Bluefish, Wellows and ourselves. They sell measurement, they price by prompts or credits, and their engine list is the product.

Tools that already sold something else and added AI. Semrush, Ahrefs, Surfer, Writesonic and Nightwatch. They price by the old unit, keywords or documents or articles, and the AI half arrives inside a plan you were buying anyway. That is why Nightwatch’s per-answer cost is misleading and why Surfer’s €49 plan has no tracker at all: the AI feature is a retention argument, not a product.

And now infrastructure. Cloudflare announced an AEO dashboard yesterday that measures from the network layer rather than by sampling, which is neither of the above and cannot be priced by prompts because it does not run any.

The buying question that falls out of this is simpler than any feature table. Ask what the vendor was selling before, because that is what the pricing is still shaped around. A tool that priced by keywords will give you a generous keyword allowance and a thin prompt allowance, whatever its marketing says about AI, and Nightwatch’s 500 keywords against 50 prompts is exactly that shape.

Where Nightwatch is the right call

You already need rank tracking, or you run an agency with many client sites. At €399 for 100 websites, 7,500 daily keywords and 500 AI prompts, with unlimited seats and white-label reports, the packaging is aimed squarely at agencies and priced accordingly.

The unlimited seats matter more than they look. Several tools in this comparison meter users, which quietly decides how many people in a company ever see the numbers.

Where it is the wrong call

You want depth on a small prompt set. Thirty answers a month per prompt is the ceiling at any price, and if you need to detect a five-point move on a specific question, no Nightwatch tier will do it. Buy something that sells repetition instead.

It is also wrong if you have no use for the SEO half, because that is most of what you are paying for.

What we could not verify

• How the thirty answers are distributed across engines and across the month. Our per-engine figures assume an even spread.
• Whether AI Overviews on Google, YouTube and Bing count as one engine or three in the response budget.
• Whether prompts run on a schedule or on demand. The page states monthly totals, not a cadence.
• What an SEO Agent credit buys, which is a separate currency on the same page.
• Enterprise pricing.

Nightwatch versus the alternatives

Against a pure AI visibility tool, Nightwatch is a rank tracker with AI included, and the honest framing is that you are buying the tracker. Against other rank trackers that added AI, it is one of the few that publishes enough for you to check what the AI half is actually doing.

The thing it deserves credit for is the same thing that produces its main limitation. Publishing prompts and responses side by side is what let us calculate thirty, and thirty is a modest number. A vendor that published nothing would have looked better on this page and told you less.

Every price here sits alongside the rest, each read from the vendor’s own page, in our comparison of every tool we verified.

Common Questions About Nightwatch

How much does Nightwatch cost?

€79 a month for Starter, €159 for Professional and €399 for Agency, with 20% off for annual billing and Enterprise at custom pricing. Every plan carries a 14-day free trial with no credit card required.

Does Nightwatch charge extra for AI visibility tracking?

No. AI tracking is included on every tier at no extra cost, which is not true of every rank tracker that has added it.

How many times does Nightwatch check each prompt?

It does not say, and it publishes the numbers to work it out. Responses divided by prompts is thirty per prompt per month on every tier. Across the four named engines that is roughly weekly per engine; across all seven named surfaces it is closer to fortnightly.

Which AI engines does Nightwatch track?

ChatGPT, Claude, Gemini and Perplexity, plus AI Overviews on Google, YouTube and Bing, and the list is identical on every plan including the cheapest. Six vendors in our catalogue put Claude on the plan you can buy and four charge more for it, so an entry-plan Claude is common rather than rare; what is unusual here is that the list never changes as you climb.

Is thirty answers a month enough?

For a single prompt, no. At that sample size a 30% mention rate carries a confidence interval of roughly plus or minus 33 points, so month-to-month movement on one question is mostly noise. Across a full prompt set it aggregates into something usable, which is how the number should be read.

Is Nightwatch good value?

If you need rank tracking, yes, because the AI half costs nothing extra. If you only want AI visibility, you are paying for 500 daily keyword checks you will not use, and tools that sell measurement alone give you more answers per euro.