The SEO content pipeline that shows its work

Audit, keywords, briefs, copy, images — one project, one thread.

  • Every fact traced to a source.
  • Every API cost on screen.
  • Every expensive step waiting for your go-ahead.

The problem

Eleven copy-pastes between twelve tools

You have a tool for every step. The work is lost between them. Context — who this company is, how it speaks, what you already published — doesn't survive the hop from a spreadsheet to a chat window. Every step starts from nothing.

Confident text, unknown provenance

Generated copy reads the same whether it rests on evidence or on nothing at all. A five-year warranty the model invented looks exactly like a real one — and goes out on your client's site as a promise to their customers.

How it works · 9 steps

From an empty project to published pages

  1. 01

    Audit

    Robots, sitemap, redirects, meta tags, schema, validator. 100+ checks, prioritised.

    critical 7warnings 19passed 86
  2. 02

    Project brief

    Company facts pulled from your site, your analytics and your notes. Each one labelled by where it came from.

    Founded 2011verified
    Engagement rings focusinferred
    Tone: warm, preciseyours
  3. 03

    Keywords

    Search volumes, two-stage cleanup, commercial and informational intent, clustering by live search results.

    collected10 846
    after cleanup2 304
    commercial · info1 412 · 892
    clusters86
  4. 04

    Pages

    Clusters matched to the pages you already have. Gaps become planned pages.

    matched to existing31
    planned pages12
  5. 05

    Content plan

    Not a list of topics — a view of your site sorted by what costs you most to leave undone.

    /rings/engagementWrite the copy
    /repairBuild the spec
    /rings/eternityCreate the page
  6. 06

    Meta tags

    H1, Title and Description in bulk. Length, keyword placement and duplicates checked automatically.

    title58 / 60
    description152 / 160
    duplicates0
  7. 07

    Page briefs

    Built from the measured top of the real SERP: length, headings, tables, FAQ, media.

    target length650–900 words
    FAQ · in 8 of 10required
    table · in 2 of 10optional
  8. 08

    Copy and images manual start

    Written to the approved brief, in your tone, linked to your real pages. Image slots generated in one pass.

    briefapproved
    /rings/engagement
  9. 09

    Analytics

    Traffic, impressions, clicks, positions — and what to fix next month.

    ━ impressions ━ clicks5 to fix

Interface fragments on this page use sample data.

What makes it different · 4 pillars

One principle: the machine must not stay quiet about what it doesn't know.

Evidence, not vibes

Every statement in the project brief carries its origin: confirmed by a primary source, inferred by the model, or told to us by you. Commercial promises — warranties, delivery times, prices, licences — are only used when confirmed. Unconfirmed claims don't get written. They become questions for the client.

DeliveryShips in 3 working dayssite · /deliveryverified
Warranty5-year warrantymodel · no sourceask client

Measured, not assumed

Target length comes from how the real top of your niche is distributed, not from a blog post about optimal word count. A jewellery workshop has a short, uniform top where a longread loses to a gallery. A medical query has a tenfold spread where completeness wins. A block is "required" because most of the top has it — not because three pages did.

Jewellery workshoptop 10 · words
0150030004500
Medical querytop 10 · tenfold spread
0150030004500

Shaded: target length for the brief.

Gates on the expensive steps

Writing is triggered by hand, one page at a time, after you've read the brief. Cheap steps run freely; costly ones wait. That is what makes the machine steerable instead of a token invoice generator.

Keywords · clusteringruns freely
Page brief · /rings/engagementapproved
Copy · /rings/engagement

Nothing fails quietly

You see how many competitor pages were read, how many clusters were covered, how many paid API calls went out and how many were served from cache, which claims were dropped and why. "No results" always comes with a reason. Invented URLs are removed by the program, not left to the model's good intentions.

Run reportValue
Competitor pages read9 / 10
Clusters covered41 / 44
Paid API calls · from cache142 · 611
Claims dropped · no source3
Invented URLs removed2

Capabilities · 11

Working today

  • Site audit
  • Project brief with sourced facts
  • Keyword research and clustering
  • Page mapping
  • Content plan
  • Meta tags in bulk
  • SERP-measured page briefs
  • Copy with real internal links
  • Image generation
  • Analytics
  • Export to WordPress, Google Drive or files
Search engines

Keyword data and SERP clustering currently run on Yandex — built for Russian-language markets. Everything else is search-engine agnostic. Google keyword data is on the roadmap.

Honest limits

What Conveyro doesn't do

Full automation with no human in the loop.

Not planned. The gates are deliberate.

Link building and outreach.

Out of scope.

Publishing anywhere.

WordPress drafts, Google Drive and file export today.

Google Search Console and GA4.

Built, pending Google's app verification.

Who it's for

For SEO specialists and small agencies running content-led SEO across several projects, tired of carrying data between spreadsheets, documents and chat windows.

Not a keyword-to-article generator. There are more steps here and more control — a deliberate price for work you can hand to a client without rewriting it.

See it on your own site

Conveyro is in private beta. Leave your email and we'll get in touch when a seat opens.

How we handle your email: privacy policy.