Your pricing, readable by the machines that recommend you
Your next customer asks an AI what you cost, and right now it's guessing. pricing.md is the machine-readable twin of your pricing page, engineered by pricing scientists so ChatGPT, Claude, Perplexity, and Gemini quote you correctly.
--- product: Acme Analytics currency: USD updated: 2026-06-12 --- ## Plan: Pro price: $499/month billed: monthly or annual (-20%) seats: up to 10 included ### Does NOT include - API access (see Enterprise) - SSO / SAML ## Add-on: extra seats price: $29/seat/month
Your next customer might be an AI agent
And it's already answering pricing questions about you, from whatever it scraped.
of brand mentions in AI search come from third-party sources you don't control.
AirOps, 2026
higher conversion from AI search traffic vs. Google search.
Webflow, 2026
of 7,000 companies in our database publish a structured pricing.md.
Valueships audit, 2026
Stripe's Agentic Commerce Protocol already in production with OpenAI.
Stripe, 2026
"Getting ingestion-ready product data is what determines whether you show up reliably across agent surfaces."
StripeWhat AI says about your pricing today, and you can't see any of it
Stale or invented prices
Models quote prices retired months ago, or invent plans that never existed.
Wrong plan for the buyer
An agency persona gets recommended your solo tier. The buyer trusts the AI, not your page.
Add-ons invisible
Anything not in clean HTML (add-ons, usage extensions, API fees) simply doesn't exist for an LLM.
"Contact sales" = dropped
Agents skip sales-gated tiers. No price anchor means no shortlist slot.
One file. Your pricing page, for machines.
A structured markdown file published at yourdomain.com/pricing.md: your complete pricing in the format LLMs parse most reliably.
Plans, limits, billing periods, add-ons, FAQ and honest comparisons, engineered so a model answering "what does your Pro plan cost?" cannot reasonably get it wrong.
For humans
Beautiful, interactive, JavaScript-rendered, and hard for an agent to read.
For machines
Clean, authoritative, always current. The source agents quote from.
Pricing scientists + a purpose-built model. Converters have neither.
HTML to MD converters 1 / 10
On our 10-point engineering matrix. Wide tables truncate, gated tiers vanish, and nothing says what plans don't include, which is where hallucinations are born.
ChatGPT prompts no science
Persona-plan mapping, value metrics, and what must never be published (your WTP, costs, discount floor). That's judgment, not formatting.
Auto-generated AI files net harm
One audited company broadcast 2018 promo prices and a private email to every LLM via a plugin-generated llms.txt. Their machine-readable layer worked against them.
Expertise layer human
150+ pricing projects. Our 9-element SaaS Pricing Canvas: 6 elements engineered into your file, 3 kept internal so competitors learn nothing. Ships with a 3-axis audit vs. Auth0, Resend and your category.
Technology layer machine
Our EU-grant-funded extraction model (led by Łukasz Sobański) reads pricing pages into our structured schema. The machine extracts; the consultants add what no model can. Subscribers get every improvement first.
Converters have technology and no pricing knowledge. Consultants have knowledge and no pipeline. We have both, plus a 7,000-company benchmark database.
: first mover in social listening, top of our 12-file audit
Every pricing.md ships with this audit: your file, scored against the best public references and your own category peers.
Auth0
Published pricing.md, the best public reference to date.
Resend
Minimalist pricing.md for volume-based email pricing.
Our deployment
First mover in all of social listening, ahead of Mention, Brandwatch & Meltwater.
| File audited | LLM readability | Pricing science | SaaS UX | Total / 30 |
|---|---|---|---|---|
|
9.0 | 8.5 | 8.0 | 25.5 |
| Auth0, published pricing.md | 8.0 | 5.0 | 5.0 | 18.0 |
| Resend, published pricing.md | 7.0 | 4.0 | 6.0 | 17.0 |
| Vercel, agent-resources (analog) | 8.0 | 3.0 | 4.0 | 15.0 |
| Typical HTML page, no pricing.md | 3.0 | 3.0 | 5.0 | 11.0 |
| Generic HTML to MD converter | 4.0 | 1.0 | 2.0 | 7.0 |
A complete package: built, deployed, and measured
pricing.md
Your complete pricing, Canvas-engineered, verified against your live page on delivery day.
llms.txt
Curated discovery file pointing AI tools to your authoritative sources, pricing first.
Deployment guide
Step-by-step for your dev team, tailored to your stack. Typical effort: 30-60 minutes.
Edit guide
For your content team: what to change safely when prices move, what to send to us.
Monitoring guide
For marketing: how to read test results and track AI answers continuously.
Baseline + post-deploy test
18 prompts × 4 models, run before launch and at week 8. Improvement measured, not claimed.
3-axis audit
Independent scoring of your file vs. Auth0, Resend, and your category peers.
Two years of being permanently AI-correct
Less than most companies spend on a single trade-show banner, for the channel where buying decisions are increasingly made.
What it costs
Risk-free start: if AI already answers about your pricing accurately and completely, you keep your $1,500.
The subscription includes
- Unlimited pricing updates. New tier, price increase, new add-on: tell us, we update the file, your team redeploys in minutes.
- Every technology change handled. Crawler behavior, llms.txt conventions, new agent surfaces, schema standards. Your file stays current.
- Change notes with every update. What changed, why, and what it means for how AI answers about you.
- Ongoing AI answer monitoring. We re-test how the four major models quote you and flag drift before your buyers hear it.
- Early access. As our extraction model matures, subscribers get every improvement first.
From kickoff to measured result in 8 weeks
Kickoff + baseline
Pricing access, persona/JTBD input, baseline test runs before anything ships.
We build
Extraction, Canvas layer, file engineering and the 3-axis audit. Full package delivered.
You deploy
Your dev team ships in 30-60 min; we verify within 48 hours.
Measure
Post-deploy test vs. baseline: accuracy, hallucination rate, persona match, citation share.
Subscription
Updates, monitoring and change notes. You track none of it.