Pipeline Overview
The problem this solves
Most SEO content workflows are disconnected:
- Keywords exported from one tool and scored in a spreadsheet
- Research and optimization done in another tool
- Writing done with AI in a chat window
- Publishing done by hand in WordPress
The result is lost context, duplicated work, and no reliable record of why a page exists or what state it is in.
SCS connects those steps into one terminal-first system driven by files you own.
The four stages
| Stage | Input | What happens | Output |
|---|---|---|---|
| 1. Discover | Keyword export from any SEO tool | Import, dedupe, score, record the human decision | keywords.csv |
| 2. Research | Keyword row | SERP review, competitor notes, NeuronWriter export, entity and question extraction | research/<pid>/ |
| 3. Draft | Research + approved brief | Brief is written and approved; AI draft is generated in opencode | md/<pid>--<slug>.md |
| 4. Publish & measure | Approved draft | Draft, pending, schedule, or publish from the terminal; review visually in WordPress; pull GSC data | publishing.csv, performance.csv |
SCS does not call SEO APIs. Data enters as CSV or JSON exports from tools you already pay for. See Data Sources and Tools.
Pipeline status flow
Selected (human decision) and QA (WordPress review) are internal gates inside
that chain. The full state machine is in Workflow.
Sites and workspaces
SCS is site- and workspace-agnostic. Any owned site, client site, or project can
be registered with scs add-site and given a PID prefix. Content is organized
per workspace under data/<workspace>/.
config/sites.csv is the only source of truth for registered sites. Run
scs sites to list them. Do not maintain a site list in these docs.
- A site is a target domain and its WordPress connection.
- A workspace is a content program. It can contain rows for one or more sites.
- A PID is the stable content identity:
<PREFIX>-<NNNN>.
What SCS owns and what WordPress owns
| Concern | SCS (terminal) | WordPress |
|---|---|---|
| Keyword and decision data | ✅ | — |
| Research evidence | ✅ | — |
| Brief and article Markdown | ✅ | — |
| Create/update drafts; read WordPress status | ✅ | — |
| Pending / scheduled / published transitions | — | ✅ via the site's approved workflow |
| Visual review of the rendered post | — | ✅ |
| Theme, plugins, taxonomy terms | config only | ✅ |
WordPress is the review surface, not the publishing control plane. Do not use the WordPress editor to add routine images, callouts, or links; put them in Markdown before publishing. See Publishing Workflow.
Next steps
- Documentation Map — pick a reading path.
- Quick Start — register a site and create a workspace.
- Keyword Research — start the method.