pSEO Engine
A programmatic SEO run usually dies in the handoffs. This one doesn't: it researches, writes, ships to a domain you control, and tells you where every page landed.
No card needed. 75 free actions — one full research run, on us. Nothing auto-charges.
Four steps. Nothing publishes without you reading it first.
Each step is one you can stop at. The research plan is yours to read before you spend actions generating anything, and the generated set is yours to read before it reaches a live domain.
Most tools collapse that into a single button. It is also what makes programmatic SEO hard to sell to a client — you cannot show them what went out, because you never saw it either.
Mine the query space, qualify against live search data, and structure the result into a page plan.
Build the page template in the design studio so the set matches the client's existing site.
Pages are written from the template and the research row behind them. You read them before anything ships.
Custom domain, generated sitemap, scheduled rank checks, weekly client report.
Every metered action has a fixed action price, listed here rather than buried. That is the most-asked question, and burying it is what makes it one.
AI research
A research run digs through the query space for a service and a market, tests each candidate against live search results, and hands you the survivors as a page plan. Queries that could only ever produce a thin page are dropped before you see them.
50 AI Actions per research run, flat
Generation and the design studio
Every page in the set grows out of the research row behind it and the one template you designed, so the batch is consistent without being copy-paste. The design studio is where that template is built, so pages read as the client's site — not as a tool's output.
2 AI Actions per generated page · 5 per design theme
AEO and GEO
Ordinary on-page SEO is table stakes. Audits also check whether a page can be parsed, quoted and cited by AI answer engines, and apply the fixes where they are missing.
1 action per audit · 1 per GEO check
Publishing and rank tracking
Pages publish to a custom domain with a generated sitemap. Rank checks run on a schedule against the published set, and the weekly client report assembles itself.
Rank checks are included in the plan, not charged per action
Figures below come from the same plan data the pricing page renders.
AI Actions meter every call the engine makes outward — a research run, a generated or refreshed page, a design theme, an AEO or GEO check. The model tokens come from a key you plug in, so token spend and model choice stay yours. Everything structural ships with the plan: publishing, custom domains, generated sitemaps, and the weekly client report.
Rank checks are capped per plan rather than charged per action, so a month of tracking cannot quietly turn into a bill. Action Packs, if you ever need them, do not expire.
Growth AI Actions
1,000
Per month, on the Growth plan
Pages per month
~500
At 2 actions a page
Client projects
3
Growth plan; up to 10 on Scale
Free to start
75
Credits, free for 7 days
After generation
The new layer is about what happens after a draft exists: catch the risks before publication, find the next best opportunities, and give an agency evidence it can show a client.
Quality gates
The publish gate checks for overlapping targets and unsupported claims alongside the existing duplicate and human-review checks. High-risk rows stay out of the live set until they are fixed.
Opportunities
The opportunity engine turns your researched rows and stored rank data into a short queue: commercial-intent pages, near-page-one refreshes, and pages seen in search but not clicked.
Client proof
Completed audits can produce a branded PDF and a useful checklist. Case-study drafts use stored pages, clicks, impressions and rankings only — no invented traffic or revenue claims.
Agency control
Funnel analytics, onboarding steps, experiments, referrals and optional client branding give agencies a clearer operating layer around every project, not just a page generator.
Want to see the same checks on your own site? Start with the free audit — no account or card required.
Get the free auditFor AI agents
Everything on this page is also an MCP tool. An agent authenticates with its own key, creates its own project, starts a research run or a generation pass, polls the job, and publishes — the same pipeline, driven by software instead of a person. No step in that sequence needs someone to open the dashboard first. Reads and setup cost nothing, so an integration can be built and tested before a balance is bought.
Keys are prepaid rather than subscribed, and an agent signs up with no card at all — one payment buys the pack outright, nothing recurs. Each key carries its own balance and its own daily spend cap, so a runaway loop stops at a number you set instead of at the end of your money, and every call it made is itemised on that key's ledger.
See what a key costs·Agent quickstart — connect in five minutes
What comes out
The fair question about any tool like this is whether the output is worth publishing. Here is one page, taken apart — pick a part.
Nothing publishes until you have read it. That is a separate step on purpose — anyone selling you a set-and-forget version of this is selling you a penalty.
The actual product
Everything else on this page is a diagram. These two are screenshots of the running app, taken from the client project in the case study.
The review queue, with the page auditor clear across all 500. This screen is the reason the black-hat answer above is not marketing: there is no setting that generates and publishes in one action, so every page in a set passes through here.
About the black bars
The client's page paths and keywords are redacted. His signed permission covers his name, his business and a link to his site — not his keyword list, which is the competitive asset of his business.
The fair question
It can be. Programmatic SEO and doorway-page spam are the same technique pointed at different questions, and a tool that generates at volume can serve either. So here is where the line is, risk by risk.
Near-identical pages spun across a location list, all funnelling to the same place, none of them worth landing on.
What this does about it
A page is only generated for a query the research run qualified as having real intent behind it. Combinations that would produce a thin page are rejected before generation, not after — which is why the page count you end up with is usually lower than the permutation count you started with.
Bulk-generated text published without anyone reading it, on the assumption that volume outruns quality.
What this does about it
Publishing is a separate, manual step. There is no setting that generates and publishes in one action, and that omission is deliberate — it is the feature request we decline most often.
One page rewritten five hundred ways, where the only thing that changes is the place name in the heading.
What this does about it
The substance comes from the research row behind each page, not from a template with slots. If two pages have nothing genuinely different to say, that is a signal the market does not support two pages — and the run says so.
Pages hosted on a vendor's subdomain, so the equity — and the leverage — sits with the vendor.
What this does about it
Pages publish to your client's own domain, with their own sitemap. If you stop paying us, the pages and their rankings stay where they are. Nothing is held hostage.
The part no tool can do for you: deciding the pages deserve to exist. The review step is where that judgement happens, and it is yours. If you publish five hundred pages you have not read, into a market that did not have five hundred questions in it, this tool will have helped you do that faster. We would rather say so here than have you find out from a manual action.
75 AI Actions covers it plus your first pages. Nothing auto-charges; you pick a plan only when the second run is worth it.
