Key Takeaways
- The Prompting Company is a YC S25 startup that raised a $6.5M seed round and now pitches "Agent Experience": getting discovered and used by AI agents, with a Discovery product and a Usability product.
- Its core loop is suggesting prompts, generating one article per prompt, publishing to an
ai.yourdomain.comsubdomain, and measuring share of voice. That covers only part of what ChatGPT actually cites. - Published plans run from $99/month to $3,000/month. Credits are the real constraint: monitoring 100 prompts daily across four engines lands you on the top tier.
- The public proof is thin. One named case study (Hyperbrowser) leaves out prompt counts, engines, and dates.
- Teams that want crawler logs, traffic attribution, or multi-client reporting often compare it against broader platforms.
What The Prompting Company is
The Prompting Company is a Y Combinator S25 company founded by Kevin Chandra (CEO), Michelle Marcelline (COO) and Albert Purnama (CTO). The founders have done this before: Cotter was acquired by Stytch in 2021, and Typedream was acquired by beehiiv in 2024. TechCrunch covered its $6.5M seed round on October 30, 2025, with Peak XV, Base10, Kearny Jackson and Y Combinator in the cap table. The team is small, around seven people according to one reviewer.
It launched in August 2025 with a blunt pitch: get your product mentioned in ChatGPT. The homepage has since moved. It now says "Get discovered & bought by AI agents," which is a bigger and more interesting claim than "get cited."
How the product works in 2026
The site splits the offer into two halves.
Discovery
Three steps: find the questions users ask, generate content to answer them, then measure AI traffic and mentions. Per the product docs, the tool suggests roughly 10 prompts a day, writes one post per prompt, and publishes them to an ai.yourdomain.com subdomain. Reported metrics are share of voice, rankings, and AI crawler traffic. There is also a "Radar" feature for seeing what buyers ask next, plus a free AI visibility audit, both marked new.
Usability
This is the more distinctive part. You map agent workflows, surface friction (API setup problems, missing docs, unclear error messages), then fix the gaps and track progress. The homepage says it is best for devtools, infrastructure, or software products with an MCP server, CLI, API, SDK, or docs-based workflow. Developer tooling around the product includes a CLI, an API and an MCP server.
I like this direction. Most of the GEO market is still arguing about mentions, while the Usability pitch asks a harder question: when an agent tries to actually use your product, does it succeed?
What it does well
- A clear niche. SaaS and devtools companies are the best fit, and the product is built around them rather than pretending to serve everyone.
- Action over dashboards. The Gamma growth lead's testimonial on the homepage says most tools are "not actionable... just providing visibility." That is a fair criticism of a lot of trackers, and The Prompting Company does ship content rather than only charts.
- Founders with exits. Two prior acquisitions is a better signal than most seed-stage GEO startups can offer.
- Agent usability work. Mapping friction in docs, APIs and error messages is useful whether or not you ever care about ChatGPT mentions.
- A lower entry price than it used to have. Earlier reviews cited $1,500/month. The October 2026 pricing page starts at $99.
Pricing and the credit math
Prices below come from the pricing page as checked by AI SEO Tracker on October 2, 2026. Confirm at promptingcompany.com before buying.
| Plan | Price/month | Includes | Engines |
|---|---|---|---|
| Basic | $99 | 25 prompts, 1,000 credits | ChatGPT, Perplexity, Gemini |
| Pro | $299 | 100 prompts, 4,000 credits, 8 articles | Adds Google AI |
| White-glove | $3,000 | 12,000 credits, 50 articles | Adds Claude, DeepSeek |
| Enterprise | Custom | Not stated | Not stated |
One response costs one credit and one article costs 20 credits, so articles come out of the same pool as monitoring. The company's COO has said they measure with about 100 prompts, daily, on four engines. That is 100 x 4 x 30 = 12,000 credits a month, which happens to be the $3,000 plan exactly.
So the cheaper tiers work out to sampling, not daily tracking. By the reviewer's estimate, Basic gives about 13 full scans a month and Pro about 10 (assuming one response means one engine answering one prompt once, which the pricing page doesn't define). That can be fine for a small team. It just isn't the same as the monitoring the company itself recommends.
Two more caveats. The engine lists are inconsistent between the pricing table, the Basic card ("all models") and the FAQ (six engines). And the free trial length and limits aren't stated.
Older price listings still float around: $1,500/month in an earlier review and "$300-$1,500/mo per brand" from Scriptbee in January 2026. Treat those as stale.
Where teams look elsewhere
The content strategy covers only part of what gets cited
Publishing a steady stream of generated articles to a subdomain targets one content type. Promptwatch Data's report on ChatGPT citation share by content type (Sep 28 to Oct 4, 2026) shows product pages at 22.4% of classified citations and landing pages at 16.8%, so about 39% of citations go to pages describing what a company sells. Listicles are 13.6%, how-tos 8.7%, comparisons 5.9%.
For commercial prompts, product pages, landing pages, listicles and comparisons make up almost 70% of citations. For branded prompts, reviews hit 5.6%, more than three times their overall share. Blog-style posts on a subdomain are one slice of that. Your product pages, your docs, and what third parties say about you matter at least as much.
The software category tells the same story. In Promptwatch's software citation data for the same week, docs and help centers earned 2.64% of software citations and vendor sites 2.01%, while review sites and analysts (G2, Gartner, Capterra and Software Advice) together earned 1.57%, more than any single vendor. GitHub led individual domains at 1.2%. If you only publish articles, you are ignoring G2 profiles and crawlable docs.
Freshness is a treadmill
Scriptbee argues that citations fade quickly and teams must keep publishing. That is a competitor talking, but the direction holds up. Promptwatch's content half-life data tracked pages ChatGPT cited on July 6, 2026 for 90 days: news lost half its pages within 36 days, while landing pages and docs lasted far longer. Volume publishing helps only if the pages you publish are the durable kind.
Proof is thin
The one public case study is Hyperbrowser, with share of voice moving from 3% to 58%, a #2 category rank and "7 of 7 categories." The reviewer notes that "+55%" is really percentage points, which is roughly 19x in relative terms. The page doesn't say how many prompts, which engines, how many runs, what dates, or what the seven categories were. The homepage logos (NVIDIA, Google, Cloudflare, Vercel and others) are logos, not case studies. Named customers from the same reviewer include Rippling, Rho, Motion, Vapi, Fondo, Kernel and Traceloop.
I'd ask for a methodology walkthrough on a sales call before trusting any single number.
Multi-brand and agency work
Scriptbee's January 2026 critique says the model suits a single brand with an engineering team rather than agencies running many clients, and claims its own product is 70% cheaper. That figure is unverified and Scriptbee sells a competing product. Still, credit pooling and a per-brand subdomain approach do make multi-client economics worth modeling before you commit.
The self-visibility question
In August 2025, a week after launch, AI SEO Tracker asked four assistants about the company. ChatGPT, Perplexity and Claude didn't know it, and Gemini returned a different company. That was a launch-week test by an author who sells a competing $49 report, so it is stale and biased. Re-run it yourself today before repeating it. I'd treat it as a reasonable thing to check, not a verdict.
Alternatives worth comparing
Product Hunt lists findable., Writesonic, PromptSignal and Citable as alternatives. G2's alternatives page leads with Birdeye, which looks like a category mismatch. I did not find independent hands-on head-to-head pricing for Profound, Peec or Otterly, so verify those yourself.
Promptwatch is the option I would compare first if you want measurement and action in one place. It tracks the real interfaces of ChatGPT, Google AI Overviews, AI Mode, Gemini, Perplexity and Copilot, and it adds IP-verified AI crawler logs, visitor analytics that tie AI traffic to conversions, and a Content Agent that publishes to WordPress, Webflow or Framer. It also watches Reddit, YouTube and LinkedIn citations, and tracks coding agents like Codex and Claude Code, which matters for devtools. Plans start at $95/month (Essential) and the Business plan at $579/month includes the automated Content Agent. Confirm at promptwatch.com/pricing.
Other tools in our catalog worth a look:
Profound
Otterly.AI
| Tool | Strength | Gap to check | Best for |
|---|---|---|---|
| The Prompting Company | Content publishing plus agent usability work | Credit limits, thin case studies, subdomain content | SaaS and devtools with docs and APIs |
| Promptwatch | Tracking, crawler logs, traffic attribution, content agent in one platform | Broader than some small teams need | Teams that want to measure revenue impact |
| Profound | Enterprise monitoring depth | Higher price point | Large brands |
| Peec AI / Otterly.AI | Simple prompt tracking | Monitoring-centered | Small teams starting out |
| Scriptbee | Unlimited domains, crawler monitoring | Vendor-reported savings unverified | Agencies with many clients |
For wider browsing, the GEO software directory at bestgeosoftware.com lists far more options by category.
Who should buy it
Buy The Prompting Company if you are a devtools or infrastructure company with a real API, CLI, MCP server or docs workflow, and you want help making agents succeed with it. The Usability work is the part competitors rarely do.
Look elsewhere if you need daily monitoring at a small-team price, per-page traffic attribution, crawler-level evidence of why you are or aren't cited, or economics that work across many client brands.
Questions to ask before signing
- What exactly counts as one response, and which engines does my plan actually cover?
- How many prompts and runs sit behind your best case study, and over what dates?
- What share of generated articles end up cited, and where do they rank among the sources?
- Can I get product and landing page work, not just subdomain posts?
- What happens to published content if I cancel?
- How long is the free trial and what is capped during it?
The vendor that answers these cleanly is usually the one worth paying.
Bottom line
The Prompting Company is a credible, founder-led product with a genuinely interesting agent-usability angle and a more reasonable entry price than it had a year ago. Its weak spots are credit-limited monitoring, a content approach centered on subdomain articles, and public evidence that is still sparse. Test it with a narrow pilot against one other platform, measure citations and AI-referred traffic over 60 days, and let the numbers decide.
If you would rather have a team run the strategy for you, 1001 SEO Media does GEO consulting and uses Promptwatch for AI visibility work.