Key takeaways
- Promptwatch publishes its numbers: Essential is $95/mo for 6,000 responses (about 1.6 cents each), Professional is $245/mo for 18,000 (about 1.4 cents), and Business is $579/mo for 42,000 (about 1.4 cents).
- Evertune's public pricing is $60/mo for 5,000 prompt credits (about 1.2 cents per credit) and $800/mo for 100,000 credits (0.8 cents). But a credit is one model sample, not one response, so the comparison is loose.
- Profound's current pricing page shows only a 7-day free trial and a custom Enterprise plan. There is no public per-response price.
- Brandlight is enterprise-only, starts with a paid 3-month pilot, and publishes no per-response figure at all.
- A tracked response is not a citation. One response usually carries 5 to 10 citations, so cost per response and cost per citation differ by that factor.
Why cost per tracked response is the number to compute
AI visibility tools love to advertise prompts. Fifty prompts sounds like a lot until you learn it runs daily across five engines in three countries. What you are really buying is volume: how many times the tool actually asks an AI engine a question and records the answer.
So the cleanest comparison is simple. Divide the monthly price by the number of AI responses (or equivalent units) included. Then check what is missing from the bundle: engines, crawler logs, content creation, add-ons.
That sounds easy. In practice, only one of these four vendors publishes a number you can divide by without squinting. I'll walk through each, then be honest about where the math breaks.
The quick comparison
| Platform | Entry price | What the price buys | Approx. cost per unit | Public pricing? |
|---|---|---|---|---|
| Promptwatch | $95/mo (Essential) | 50 prompts, 6,000 responses, all models | ~$0.016 per response | Yes, all tiers |
| Promptwatch | $245/mo (Professional) | 150 prompts, 18,000 responses | ~$0.0136 per response | Yes |
| Promptwatch | $579/mo (Business) | 350 prompts, 42,000 responses | ~$0.0138 per response | Yes |
| Evertune | $60/mo (Starter) | 5,000 prompt credits, up to 5 models | ~$0.012 per credit | Yes |
| Evertune | $800/mo (Pro, annual) | 100,000 credits, 11 models | $0.008 per credit | Yes |
| Profound | Free 7-day trial, then custom | Trial: 50 prompts daily on 3 engines | Not published | Enterprise is quote-only |
| Brandlight | Custom, paid 3-month pilot | Enterprise scope | Not published | No |
The per-unit figures for Promptwatch Professional and Business and for both Evertune tiers are straight division from the published plan details. Confirm everything against each vendor's live pricing page before you budget, because these pages change.
Promptwatch: the most transparent math
Promptwatch is the easiest of the four to price out, because response counts are printed on the plan.
| Plan | Monthly price | Prompts | Responses per month | Cost per response |
|---|---|---|---|---|
| Essential | $95 | 50 | 6,000 | ~$0.0158 |
| Professional | $245 | 150 | 18,000 | ~$0.0136 |
| Business | $579 | 350 | 42,000 | ~$0.0138 |
A few things worth noticing.
The per-response price barely moves between Professional and Business. You don't get a big volume discount for stepping up. You step up because Business adds the automated Content Agent, state and city tracking, local GEO/SEO targeting at exact coordinates, Shopping Insights, Ads Radar, and a bigger crawler log allowance (100M versus 25M).
The bundle matters more than the headline rate. Every paid plan unlocks all active models, so you are not paying extra to add an engine. Essential also includes country tracking, a site audit, 360 Insights, content gap analysis, and MCP and API access. Responses are collected from the real interfaces of ChatGPT, Google AI Overviews, Gemini, Perplexity and others, not only from APIs.
Annual billing gives 12 months for the price of 10, so the effective rate drops by roughly 17%. Essential and Professional also come with a 7-day free trial on monthly billing. One caveat: at least one third-party blog quotes Professional at $249/mo. Promptwatch's own number is $245, so trust the pricing page at promptwatch.com/pricing.
Agencies get a separate ladder. Kick-off is $199/mo for 10,000 responses (about 2 cents each), Growth is $399/mo for 25,000 (about 1.6 cents), and Scale is $799/mo for 65,000 (about 1.2 cents), with unlimited projects and prompts on all three.
Evertune: cheap per credit, but a credit is not a response
Evertune's pricing page is public, which already puts it ahead of two of the other three vendors.
Starter is $60/mo for 5,000 prompt credits across up to 5 models (ChatGPT Search, Gemini Search, AI Mode, AI Overview and Copilot). Pro is $800/mo, billed annually, for 100,000 credits across 11 models, adding ChatGPT Base, Gemini Base, Meta AI, Claude, Perplexity and DeepSeek. Pro also adds a Content Agent with 25 articles a month, Prompt Lab, Prompt Volumes, data download and a data MCP. Enterprise is custom and adds bot analytics, a site audit, SSO and cloud/API export.
The arithmetic: $60 / 5,000 is $0.012 per credit, and $800 / 100,000 is $0.008.
Here is the catch. Evertune's credit is defined as one prompt sample on one model. Its calculator uses prompts x models x samples per day x 30 days. Fifty prompts on one model at three samples a day is 4,500 credits, which just fits Starter. Evertune also leans on statistical significance, meaning around 100 samples per prompt. A single prompt on a single model can burn 100 credits.
So is a credit the same as a Promptwatch response? Roughly, if you treat each sample as one response. But that is my assumption, not a vendor-confirmed equivalence. Treat the 0.8 to 1.2 cent figure as a rough lower bound, and note that Starter does not include the day-one significance sampling that Pro does. Also note the Ads Agent is priced separately at 10% of media spend on Starter and 8% on Pro.
Profound: no public per-response price
Profound is the best-known name in the category and the hardest to price.
The current pricing page lists two options. The free trial is limited: 50 prompts run daily for 7 days on ChatGPT, Gemini and Google AI Overviews, in one language and one region, with a recommended prompt set you cannot customize, and no exports, API or data history. The Enterprise plan is custom, with a tailored prompt plan, up to 9 answer engines, ChatGPT Shopping, CSV/JSON export, API, all-time history, SSO/SAML, SOC 2, unlimited seats and Agent Analytics.
You may see older figures in the wild: $99 Starter (ChatGPT only) and $399 Growth (three engines, 100 prompts, 6 articles). Several third-party roundups, and Brandlight's own comparison, still cite them. They are not on the current pricing page. Even Profound's own June 2026 comparison content has quoted Growth at roughly $499. Treat all of these as stale or unverified.
The only public unit you can compute is the trial: 50 prompts x 3 engines x 7 days is 1,050 responses, free. Beyond that, any cost per response is guesswork until you get a quote. Agents also run on credits (the FAQ mentions 400 credits per month per client workspace on the Agency Growth plan), and overage can be billed or paused depending on account setup, so factor that in.
The tradeoff is obvious. You get a deep enterprise feature set and likely negotiated volume, but you can't know what a response costs until you are in a sales conversation.
Brandlight: priced to scope
Brandlight says openly that it is enterprise-only and priced to scope, starting with a paid 3-month pilot rather than a free trial.
There is no per-response number to divide. A third-party tracker, IndustryLens, reports enterprise tiers reaching up to $25K per month, but that is unverified and I would not budget off it. What it does tell you is that Brandlight is not competing with a $95 or $60 starter plan. It is selling a managed, enterprise engagement that tries to cover the third-party sources AI engines cite (its own analysis claims roughly 85% of cited sources for category questions are third-party or social).
If you are evaluating Brandlight, cost per response is the wrong first question. Ask what the pilot covers, how many prompts and engines are in scope, and what you will be paying after the three months.
Why the units do not line up
Three different things hide behind the word "tracked":
- A response is one answer from one engine to one prompt. Promptwatch bills on this.
- A sample is one run of a prompt on a model. Evertune's credits count this, and its statistical approach uses many per prompt.
- A citation is one source link inside a response. According to Promptwatch Data on average sources per response, ChatGPT cites about 5 sources per web-search answer, Google AI Overviews about 10, and Perplexity almost exactly 10 every day. Copilot has swung from under 2 to nearly 17.
So if you care about cost per citation observed, multiply the response count by the citations per answer. A 6,000-response Essential plan could yield tens of thousands of citation data points. This is also why a tool with few responses but a deep citation breakdown might beat a cheaper one with shallow output.
One more wrinkle. ChatGPT's own behavior shifts, and that changes what a response contains. Promptwatch's data shows ChatGPT started using the site: operator at scale on August 8, 2026, with searches per response nearly doubling. Tools that measure fan-outs capture this; tools that only count top-line answers do not.
What the cheap rate leaves out
A low cost per response is worthless if you must buy five other tools to use it. Here is what each vendor bundles, based on the research above.
| Capability | Promptwatch | Evertune | Profound | Brandlight |
|---|---|---|---|---|
| Self-serve plans | Yes | Yes | Trial only | No (pilot) |
| All models on paid plans | Yes | Starter 5, Pro 11 | Trial 3, Enterprise up to 9 | Scoped |
| AI crawler logs | Yes, from Professional | Enterprise bot analytics | Agent Analytics (Enterprise) | Not confirmed |
| Visitor and conversion tracking | Every plan | Not confirmed | Yes | Not confirmed |
| Content creation | Content Agent, Business and up | Content Agent on Pro | AI Marketer credits | Not confirmed |
| Published per-response price | Yes | Per credit | No | No |
Promptwatch's own comparison of seven GEO tools makes the same point about its entry plan: crawler logs and human visitor tracking together on self-serve plans are rare. Of course, that comparison was written by Promptwatch's team, so weigh it accordingly. I'd verify the crawler log claim during a trial rather than take anyone's word for it.
How to pick based on budget
You want predictable spend under $300 a month
Promptwatch Essential or Professional, or Evertune Starter. Evertune is cheaper per credit but tracks fewer models and skips the significance sampling. Promptwatch costs more per unit and includes all models, site audit, content gap analysis and API access from day one. The free Explore plan (10 prompts, ChatGPT only) is a reasonable way to test before paying.
You need statistical rigor on a few high-value prompts
Evertune Pro makes sense here. At 100 samples per prompt, 100,000 credits covers roughly 1,000 prompt-model pairs, and the Content Agent is bundled. Just run the credit calculator on your own prompt list before committing to $800/mo annual.
You are an agency
Promptwatch's Agency plans start at $199/mo with unlimited projects and prompts, white-label reporting and branded client access. Profound also has an Agency Growth plan, but the page gives no price.
You are an enterprise with a security review
Profound (SOC 2, SSO/SAML) and Promptwatch (ISO/IEC 27001, GDPR, SSO on Enterprise) are the two to put side by side. Brandlight belongs on the list if you want a managed engagement. Ask all three for a written quote that states responses or samples per month, engines, regions and overage terms.
Questions to ask before signing
- What is the unit: response, sample, prompt, or credit?
- How many engines are included, and are they scraped from the real UI or called through an API?
- Is the region and language count included or an add-on?
- What happens when I exceed my allowance: overage fees, a pause, or a forced upgrade?
- Are crawler logs, visitor analytics and content tools in the plan or priced separately?
- Can I see a current pricing page, in writing, dated?
If a vendor cannot answer the first question, you cannot compare costs. Full stop.
Where to look next
If you want a wider field than these four, the directory at bestgeosoftware.com lists GEO platforms by category, and Promptwatch Data publishes the citation research behind some of the numbers above.
Bottom line
On published numbers, Evertune has the lowest per-credit rate, Promptwatch has the clearest per-response rate (about 1.4 to 1.6 cents) with all models included, and Profound and Brandlight cannot be compared until you ask for a quote. Pricing pages change often, so confirm each figure on the vendor's site before you put it in a budget.