Written for how AI engines actually read
Answer engines split a page into passages and cite one of them. ActiveGeo writes to that unit, scores it against 15 checks you can audit, and checks whether the engines can reach your site at all.
Written to be cited
AI engines split a page into passages and quote one of them. That is the unit we write for.
Passage-level templates
Ten GEO templates, each with a contract for every retrievable passage — what it must open with, what it must contain, how many concrete figures it needs. Definitive Answer, X vs Y, Best-of, How-to, Question Cluster, Statistics, Troubleshooting, Pricing, Data Study and Entity Profile.
Answer-first, self-contained sections
Every section states its answer in the first two sentences and names its own subject, so a passage lifted out of the page still makes sense. Around 44–55% of AI citations come from the top of a page, in two independent analyses.
Real sources, never invented
Citations are mandatory, not optional — and only to URLs from verified research. Any link the model invents is stripped before you ever see it. It will not fabricate a statistic, a price or a quote to satisfy a check.
A GEO score you can audit
Fifteen deterministic checks across extractability, information gain, query match and machine readability. Every check shows its rule and the study behind it — no black box. Below the target, one targeted revision fixes only what failed.
See what engines see
A page an AI crawler cannot fetch cannot be cited, however well it is written.
Chunk view
Your article as a retrieval engine reads it: one card per passage, each scored on its own, each flagged if it opens with a pronoun, buries its answer or carries no concrete figure.
Crawler-access diagnostics
We read your robots.txt crawler by crawler and tell you which blocks cost you citations and which cost you nothing. Blocking GPTBot is free; blocking OAI-SearchBot removes you from ChatGPT entirely. Plus rendering, snippet directives and speed.
AI traffic from your server logs
AI crawlers do not run JavaScript, so no analytics snippet can see them. Add one middleware file or paste an access log, and see which crawlers came, which pages they never fetched, and which of your URLs a live-fetch agent pulled in response to someone asking a question.
Entity and internal linking
Automatic internal links to your own published posts using real, resolvable URLs only — plus Organization and author identity with sameAs, which is what tells engines which brand you actually are.
Still fully SEO-complete
GEO is roughly 70% classical SEO done properly. We did not throw that away.
Deterministic SEO engine
Thirteen on-page checks scored 0–100 against your target keyword — keyword placement, heading structure, title and meta length, alt text, internal and external links, readability. Shown beside the GEO score, never merged with it: the two sometimes disagree.
Live AI editor
The draft streams into a clean editor as it is written. Highlight any line to rewrite, expand, shorten, change tone — or make a passage stand alone.
Automatic illustration
A hero image and section visuals sourced and placed automatically, skipping non-visual sections, deduped and capped. Off by default in GEO mode, because page speed correlates with citation and stock imagery does not.
Any model, or your own key
Switch between Claude, GPT, Gemini and DeepSeek per task. Or bring your own provider key and pay your provider directly with no markup.
Publish without losing it
Most GEO work is thrown away at publish time. This is where it survives.
Schema on every destination
Article, author Person and Organization JSON-LD with sameAs, cited sources and an accurate dateModified — to your blog, to WordPress, or to any webhook receiver. Not just the destinations that are easy.
Canonical across destinations
Publish one article to three places and you get three competing URLs. The first destination owns the canonical; the rest declare it, so the signal concentrates instead of splitting.
Automations with a quality gate
Schedules rotate through your keywords on your cadence — and can hold any post scoring below your bar in the review queue instead of publishing it. Autopilot should not ship content that fails your own standard.
Content API + signed webhooks
A REST API to generate content, manage templates and drive schedules. Webhooks carry the finished article plus ready-made schema, cited sources, entities and canonical — with retries and a delivery log.
Scale and control
Publish on a schedule, unattended
Pick your topics once. On the days you choose we research, write and file a full article to your site — as a draft you approve, or live. It rotates topics so no two posts repeat.
A content plan you can see
The planner shows what is queued, drafted and published across the month, per site — so a gap in the calendar is visible before it is a gap in your traffic.
Your own templates
Describe what you want and AI drafts a reusable template — including its own passage plan. Private to you, reused across the editor, automations and API.
Real-euro cost control
Usage metered in actual euros with a reserve-then-reconcile flow, so failed generations cost almost nothing and you always see the true cost per article.
Doc-aware chat assistant
A built-in assistant that reads your draft. Ask it to add a section, tighten a paragraph or answer a question, and apply its output straight into the editor.
One unified composer
A single workspace instead of scattered tools. Guided flow with an editable outline, or a fast one-shot article.
What we deliberately don’t build
The GEO market is full of things that sound right and don’t survive testing. We leave them out and say why.
- llms.txt generation — Tested across ~300,000 domains with no measurable effect on citations. Google compared it to the meta keywords tag.
- Schema sold as a citation lever — A controlled study of 1,885 pages found roughly zero lift. We emit schema for entity identity and agent readability, and say so.
- A single blended “AI visibility score” — Only 3.8% of sources are cited by all four major engines. One number hides more than it shows.
- Guaranteed citation claims — Only about 10.6% of cited URLs are still cited 28 days later. Anyone promising a stable position is selling noise.