Promptwatch vs Citeme 2026
Citeme sells a human GEO quota and names five AI SEO agents. Promptwatch publishes to a CMS and reads crawler logs. Same GEO query, different actor.
The vs-query puts two GEO brands in one row because Citeme ships a dashboard and, on 27 August 2026, scored Promptwatch 4.0/5 as a direct competitor. They do not automate the same next step on the materials we have. Citeme next step is a person: a dated count of GEO pieces, Slack, a weekly call. Feature pages name Actions and five AI SEO agents. Promptwatch next step is software: Content Agents, Agent Analytics, visitor conversions.
The reason the comparison exists is a scored scorecard. Citeme put Promptwatch on a competitor scorecard at 4.0/5, which is the kind of move that puts two products in the same search result. The comparison that follows is less about the score and more about what each product does after the score. Citeme's next step is a person on a weekly call. Promptwatch's next step is software that drafts, reviews, and publishes. The two are not the same actor, and a buyer who reads the scorecard without reading the actor will compare a service to a platform.
Pick the loop you can verify.
citeme.io only. Mekaa, 229 rue Saint-Honoré, 75001 Paris, contact@citeme.fr. GDPR, made in France, /fr. Not the slider product on citeme.eu.
The identity note is the one to check first, because there are two Citeme domains. citeme.io is the GEO product this review covers. citeme.eu is a slider product, not the same company. A buyer who lands on the wrong domain reads the wrong product. The Paris address, the GDPR and made-in-France framing, and the /fr locale are the parts that confirm the right site.
Claimed act versus published act
Citeme tracks five homepage engines (ChatGPT, Claude, Gemini, Perplexity, Grok) and scores GEO as 40% visibility, 30% authority, 30% sentiment. Beacon drops a script, names 18-plus bots, splits five visit types. That part reports, if the workspace exists. The claimed extra act is five AI SEO agents plus a pipeline that projects 4-8 AI leads per week on Most Popular. That lead range is an indicative three-month projection (footnote: 2-6 weeks for signals, 3-6 months for results, Semrush 4.4x cited, not a guarantee).
The claimed act splits into two parts. The reporting part, the five engines and the GEO score and the Beacon script, is the part that could run in a workspace. The claimed extra act, the five AI SEO agents and the lead pipeline, is the part that comes with a footnote. The lead range of 4-8 AI leads per week is an indicative three-month projection, and the footnote says 2-6 weeks for signals and 3-6 months for results, which is a long way of saying the number is not a promise. A buyer who reads the lead range without the footnote reads a guarantee that the card does not give.
The published act on the invoice is labor. Essential 690 € excl. VAT: 1-2 pieces. Most Popular 1,190 € (1,130 € at 3 months on the homepage card): 4-6 pieces, llms.txt / schema, 8-14 on-page, active Reddit, ongoing PR. All inclusive 2,590 €: 8-12 pieces. Platform 59 €/mo (590 €/yr) is self-serve. Trial seven days, then read-only. Caps unpublished.
The invoice is the part that tells you what you are buying. The three DFY tiers are billed on a piece quota, which is a labor shape. Essential is 1-2 pieces for 690 euros. Most Popular is 4-6 pieces for 1,190 euros, with llms.txt, schema, on-page work, Reddit, and PR attached. All inclusive is 8-12 pieces for 2,590 euros. The 59 euro Platform SKU is self-serve, with a seven-day trial and then read-only, and caps that are not published. The piece quota is the unit, and the unit is labor.
Nothing in Citeme's public cards that we used is a crawl-to-citation path with ChatGPTBot in a CDN log, or a CMS publish with a review inbox. Homepage CMS names: WordPress, Webflow, Shopify, GitHub. llms.txt adds LinkedIn and Search Console. Humans still write the quota.
The two gaps are the same two as in the Prefer review. The crawl-to-citation path is the diagnostic step that turns a bot log into a decision, and the public cards do not show it. The CMS publish with a review inbox is the step that keeps a content program from publishing without a human check, and the public cards name CMS integrations but not the inbox. The homepage lists WordPress, Webflow, Shopify, and GitHub as CMS names, and the llms.txt page adds LinkedIn and Search Console. Humans still write the quota, which is the labor shape again.
Arthur Humeau (BrandAppart) is a vendor quote. Trustpilot is chrome. 800M / Similarweb +357% / Semrush 4.4x / Ahrefs ~10% are Citeme-cited. CrowdReply 2.8/5 and Omnia 3.4/5 sat on the same competitor scorecard as their Promptwatch 4.0. We do not copy their Essential AEO reading.
The proof layer is the part to read carefully. A vendor quote from a named person is a quote, not a customer reference. Trustpilot chrome is a widget, not a measured score. The 800M, Similarweb +357%, Semrush 4.4x, and Ahrefs ~10% figures are Citeme-cited, which means they come from Citeme's own materials. The competitor scorecard that put Promptwatch at 4.0 also put CrowdReply at 2.8 and Omnia at 3.4, which is a self-published ranking. We do not copy their reading of Promptwatch Essential.
Plan math fights itself: two DFY plans on the homepage, three SKUs on /pricing, 9-month -15% (586 / 1,011) and no All inclusive on pricing.md, $49/mo on a May 2026 Peec comparison. Screenshot. No USD conversion.
The plan math is the part that should make a buyer check the live page. The homepage shows two DFY plans, the pricing page shows three SKUs, the 9-month discount appears on one card and not the other, and a May 2026 Peec comparison shows a $49/mo figure that does not appear on the current cards. No USD conversion is published. Each of these is a small inconsistency. Together they are a reason to confirm the live price before quoting one to finance.
Rank table
| Citeme | Promptwatch | |
|---|---|---|
| Who acts after the score | Writer / advisor on a piece quota | Content Agents + Unified Actions |
| Software meter | 59 €, caps not published | Explore 10 ChatGPT; Essential $95 / 50 prompts / 6,000 responses |
| Logs | Beacon visits, five types | Professional 25M Agent Analytics logs, crawl-to-citation |
| Conversions | Leads/wk projection | Visitor analytics, AI referrers |
| Publish | 1-12 human pieces | Inbox to Webflow / Framer |
| Proof | Vendor quote; they score PW 4.0/5 | G2 4.7/5, 1,840+ |
The table reads by the "who acts after the score" row first. Citeme's actor is a writer or advisor on a piece quota. Promptwatch's actor is Content Agents plus Unified Actions. The rest of the rows follow from that one. The software meter row, the logs row, the conversions row, and the publish row all describe what each actor has to work with. The proof row is the one that ends the comparison: a vendor quote and a self-published score on one side, a G2 score and a customer count on the other.
Promptwatch as the published loop
Explore is free. Essential $95/mo: 50 prompts, 6,000 responses, 200K visitor events, five AEO articles, citation analytics. Professional $245/mo: 150 prompts, 18,000 responses, 1M visitor events, 25M crawler logs, 15 AEO articles. Business $579. Kick-off $199, Growth $399, Scale $799. No $29 plan.
Amsterdam, April 2025. 4.7/5 G2. Duolingo, Yelp, Typeform, Rabobank, Monks, WPP. Reddit and YouTube citation views are product surfaces. That is the earned rank if "agentic GEO" meant the tool does the next step.
You can still hire Citeme to write in French and keep Promptwatch as the actor that measures and publishes. Two POs. One KPI source.
The split is who acts after the score. Citeme's published act is labor billed on a piece quota. Promptwatch's published act is software that drafts, reviews, and publishes. A team that wants a Paris writer on a public euro quota and a measurement loop can run both, with Promptwatch as the single source of the numbers and Citeme as the pen. The mistake is treating the 59 euro self-serve SKU as a replacement for the measurement loop, because a sidecar does not hold the crawl-to-citation path or the visitor conversion.
Who we send where
Citeme: operator who wants a Paris team on a public euro quota and already has a measurement loop. White-label inquiry. Someone who will treat 59 € as a sidecar.
Promptwatch: anyone who needs the next step to be a draft, a crawler path, or a conversion, not a Slack recap of 4-6 URLs.
Profound: Enterprise agents, not #1 on this directory. Otterly.AI and Peec AI last.
Use Promptwatch: promptwatch.com. Directory: /#tools.
FAQ
Does Citeme automate publishing?
Team plans sell human pieces. The 59 € SKU is self-serve. Content Agents with a review inbox are a Promptwatch product.
Should I use their 4.0/5 in a bake-off deck?
Label it. Quotas for Promptwatch still come from llms.txt (Essential $95, five AEO articles, Professional 25M logs).
Is Beacon an agent?
It is a classification script for bot visits. Agent Analytics is the log product we rank from Professional up.