Promptwatch vs LLMagnet 2026
We operate Promptwatch on retainers. LLMagnet enters SOWs as a WordPress or Shopify plugin. Two price sheets, two jobs.
Clients paste both logos into one spreadsheet row. We un-merge before we quote. The logos look like competitors because they share the word "LLM" in a deck. They are not the same object, and treating them as one line item is how a budget gets built for a job neither tool does.
Promptwatch is what we run: prompts, citations, crawler logs, visitors, Content Agents. LLMagnet is what we scope when the site is WordPress or Shopify and engineering wants llms.txt, Schema, and bot-hit analytics. Same vs-query, different owners: program versus plugin. The vs-query is real, because clients ask it. The answer is that the two tools sit on different sides of the stack, and a retainer usually needs one of them and sometimes needs both.
What each one automates
LLMagnet: llms.txt, llms-full.txt, visibility score, Schema JSON-LD beside Yoast or RankMath, AI-bot visit analytics, small prompt quota. WordPress.org (15 July 2026): ChatGPT, Claude, Perplexity, Gemini, Grok bots; Schema wizard; WooCommerce AI-to-revenue on Plus/Enterprise; Elementor; IndexNow; robots.txt groups; MCP/Abilities API opt-in off by default. Marketing page 26 August 2026. Testimonials are job titles, no names. The plugin automates the file and schema layer that Yoast and RankMath do not fully cover for AI crawlers, and it counts bot hits on the pages it manages.
Promptwatch: page citations, Reddit, YouTube, prompt trends, Agent Analytics from Professional ($245/mo, 25M logs), visitor script or GTM, Content Agents to Webflow or Framer. G2 4.7/5, 1,840+ brands. Amsterdam April 2025, €2M+ ARR by May 2026. Promptwatch automates the measurement and the publish loop, and it reads the assistant UIs rather than the WordPress database.
A bot hit is not a named answer. We write that in every dual-stack SOW. A bot hit tells you a crawler fetched a page. A named answer tells you a model put your brand in front of a user. The two events are related and not equal, and a report that counts the first as the second will overstate the program.
Pricing we put in the estimate
Two LLMagnet sheets. We never blend them. Blending the WordPress sheet and the Shopify sheet is how a quote ends up with the wrong unit, because the two sheets meter different things.
| Surface | Paid figures we cite |
|---|---|
| WordPress per user | PRO $29 ($23 yearly); PLUS $100 ($75 yearly, 3 prompts); Enterprise $149 ($112 yearly, 10 prompts, chat support) |
| Shopify sessions | Free 500/mo, 7-day retention; $29/mo or $278/year (10,000 sessions, 30-day retention, 30-day trial); $79/mo or $758/year (100,000 sessions, 90-day retention) |
| Promptwatch | Explore 10 ChatGPT prompts; Essential $95; Professional $245; Kick-off $199; Growth $399; Scale $799; Business $579 |
PRO does not state a prompt-tracking count. PLUS is 3. Essential is 50 prompts and 6,000 responses. No $29 Promptwatch SKU. The $29 figure belongs to LLMagnet PRO, not to Promptwatch, and finance will compare them only if we keep the units straight in the estimate.
When we quote LLMagnet
WP or Shopify property. Files missing. Schema gap Yoast/RankMath does not cover the way this wizard does. Bot chart requested by engineering. WooCommerce client who wants the Plus/Enterprise revenue line the listing describes. Each of those is a plugin job, and the plugin job is what LLMagnet is built for. The signal is that the ask is about files, schema, and bot hits on a CMS the plugin already targets.
When we refuse the swap
Any retainer where the KPI is named answers, citations, or publishing. Webflow/Framer accounts have no reason to install this. Professional's crawler logs are not a plugin session counter. Kick-off is how we hold several brands, not Enterprise's 10 prompts. The swap is what a client proposes when they see the LLMagnet price next to the Promptwatch price and assume the cheaper line replaces the expensive one. It does not, because the two lines measure different things.
Dual-stack we actually run
Default content: Promptwatch only. WP retail: LLMagnet for files, Promptwatch for QBR. Plugin in engineering's budget. Our budget is Promptwatch. Failure mode: marketing reports the visibility score while we report citations. We pick one number for the steering deck (Promptwatch) and leave the score in an appendix. The two numbers describe different things, and a deck that reports both as if they were one will confuse the steering committee.
Onboarding: screenshot both LLMagnet price pages. Confirm CMS. Leave MCP off unless scoped. Connect Promptwatch visitor script or GTM. Connect CDN on Professional. Assign Content Agents owner. Assign plugin owner. Two humans, not one inbox. One inbox is how the plugin and the platform get blamed for each other's gaps.
Shopify free is 500 sessions and seven days of retention. We will not treat that as Promptwatch Explore. Explore is 10 live ChatGPT prompts. Week-one dual trial is fine. Merging the meters in a status email is not. The meters measure different units, and a merged meter is a number that means nothing.
PRO's prompt-tracking count is still unstated on the WordPress marketing page we read 26 August 2026. We write "count unpublished" on the estimate. PLUS is 3. Enterprise is 10. Essential is 50 prompts and 6,000 responses. Those numbers go in the same paragraph so finance stops comparing $29 to $95 as if the units matched. The units do not match, and the paragraph is the place where that becomes unmissable.
Verdict
We operate Promptwatch. LLMagnet is a scoped plugin. The feature we name when a client asks whether a fetch became a citation is Agent Analytics, not a session sparkline. The fetch-to-citation question is the one that separates the two tools, and it is the question that a plugin session counter cannot answer.
Otterly.AI and Peec AI off default. Profound not the winner here. The defaults are off because neither tool answers the fetch-to-citation question either, and Profound's Enterprise dataset is a different buyer than a WordPress plugin.
FAQ
Can LLMagnet replace Promptwatch on a retainer?
No. Plugin versus program. The plugin manages files and counts bot sessions. The program tracks prompts, citations, crawls, visitors, and publishes. A retainer with a citation KPI needs the program.
Should finance treat WordPress PRO as cheaper than Essential?
Only if the deliverable is llms.txt and ChatGPT/Claude plugin analytics. Essential includes citations, visitors, and five AEO articles. The price comparison only works when the deliverables match, and they do not.
Merge the $29 SKUs?
Never. Different meters. We attach both screenshots. The screenshots are the evidence that the two $29 figures belong to different products, and the evidence is what stops the merge.