An SEO department,
evidence first.
A five-seat agent team runs the audit the way a serious agency would: each seat with its own browser and its own job, every claim measured on the live SERP and cited to your data. The output is a complete, client-ready audit that your analyst reviews and signs.
First month freeThen £99 / seat / month · runs the most capable claude models
Five seats, each with its own browser.
An SEO Manager sequences the campaign, and four specialists work it in parallel. Every seat drives its own embedded browser, takes its own screenshots, reads your connected data sources, and owns one part of the audit, so nothing is asserted that was not measured.
You watch all of it live: per-agent terminals and browsers stream in the app while the campaign runs, and you can steer mid-run.
What the seats actually examine.
Eighteen methodology skills, red-teamed against the hardest question in the industry: would a named expert debunk this?
Live SERP measurement
Rankings, features and AI Overview presence are checked on the real results page at audit time, never assumed from a stale export.
Full crawls of the cluster
Indexability, canonicals, redirects and orphan pages mapped across the money cluster, blockers separated from tie-breakers.
Visual semantics
Agents screenshot your pages and the ranking set side by side and compare what a searcher actually sees, not just the markup.
Your Search Console and GA4
Clicks, impressions and conversions pulled from your connected properties, so every opportunity is sized in your own numbers.
Link-gap intersects
Referring domains that link to multiple ranking competitors but not to you, computed from Ahrefs, Semrush or DataForSEO.
Intent and cannibalisation
Every target query classified against the live SERP, and overlapping pages that split one query mapped page by page.
Topical coverage vs the leader
Your cluster mapped against how the leading competitor structures the same topic: pillar, support pages, internal links.
A reproducible evidence log
Every number records where it came from and when it was measured, so any finding can be re-run and defended later.
Built to be signed by your analyst.
The team does the weeks of legwork; your analyst does the judgment. What lands is a complete audit with severity-ranked findings, charts, and the evidence behind every number, ready to double-check and put your name on in front of a client. Nothing ships that cannot defend itself.
- Severity-ranked findings with the fix, not just the fault
- Every claim cited to its source, or marked provisional
- A human signs it: agents draft, your analyst approves
The most capable models in the world, on every task.
The team runs on your own Claude subscription, which means the frontier models do the analysis. Codus never swaps in a smaller model to protect a margin, because it has no margin to protect: you bring the account, and quality is never traded for cost.
Numbers come from your sources, cited.
Connect once in Settings. Agents route each metric to the best connected source and name it on every number. Nothing is assumed available; everything is pinged first.
Site rebuilds backed by a commissioned study.
Point codus at an existing business site. It commissions its own SEO study first: keyword maps, competitor teardown, architecture. Then it rebuilds every page against those deliverables and hands over the full research with the site. One prompt in, a research-backed site plus a client-grade report out.
Switch SEO on per seat.
The SEO discipline is a £99 per seat monthly module on top of your codus seat. Give it to the people who own search, leave it off for the rest.