Publish SearchChamp drafts from your own editor.
Pull a draft, edit it in your own repo or site, and report the publish back — so rank tracking and outcomes attribute the change correctly.
Four tool calls.
This workflow is for teams who don’t publish through the SearchChamp dashboard or the WordPress integration — you keep full control of your repo or CMS, and SearchChamp just needs to know where the content ended up.
List your draft articles
Ask Claude Code or Cursor to call list_draft_articles on the SearchChamp connector. It returns every draft in your workspace that's ready for a human (or an AI pair-programmer) to take over.
list_draft_articles(site_id)
Pull the full draft
Call get_article_content for the draft you want. It returns the full body, metadata (title, meta description, target keyword), and the JSON-LD structured data SearchChamp generated for the piece.
get_article_content(article_id)
Edit it in your own repo or CMS
Make the edit exactly like you would any other file — in your static-site repo, your CMS integration, or wherever your content actually lives. Nothing about this step is SearchChamp-specific.
Report the publish
Once the page is live, call report_publish with the article ID and the canonical URL you published it at. This is what lets rank tracking and outcome attribution find the page and start measuring it.
report_publish(article_id, canonical_url)
The whole sequence, at a glance.
list_draft_articles(site_id) → get_article_content(article_id) → (edit in your repo / CMS, then deploy) → report_publish(article_id, canonical_url)
Publishing from code — FAQ
No — this is for teams who don't publish through the SearchChamp dashboard or the WordPress plugin and want full control of their own repo or CMS instead. If you're on WordPress, the WordPress plugin publishes directly with no separate tool calls needed.
The pull-publish workflow uses the same MCP connector as the rest of the Claude and ChatGPT integrations, so it's available anywhere the connector is — no separate plan requirement beyond having a connected SearchChamp workspace.
The full article body, its metadata (title, meta description, target keyword), and the JSON-LD structured data SearchChamp generated for the piece — everything you need to publish it correctly in your own repo or CMS without hand-recreating the SEO metadata.
The page goes live, but SearchChamp never learns its canonical URL — so rank tracking and outcome attribution can't find it or start measuring it. The content itself is unaffected; you just lose the automatic connection between the draft and its published performance until you call report_publish.
Edit it exactly like you would any other file in your repo or CMS — nothing about this workflow expects the draft to be published unedited. report_publish only needs the article ID and the final canonical URL once it's live, regardless of what changed in between.
Any of them — a static-site repo, a headless CMS, or a custom publishing pipeline. SearchChamp only needs the canonical URL back via report_publish; it doesn't need to know how or where you actually built and deployed the page.
Yes — they're independent publishing paths into the same SearchChamp workspace. Some articles can go out through the dashboard or WordPress plugin, others through this code workflow, and rank tracking treats them the same way once report_publish (or the equivalent dashboard/plugin publish step) has recorded the canonical URL.
Set up the connector first.
Claude Code and Cursor use the same connector as the Claude web and desktop apps — add the connector URL from Settings → Connect your AI apps once, and every MCP-compatible client on your machine can reach it. If you hit an authentication error partway through, the troubleshooting guide covers OAuth loops and workspace scoping.