Combine Read a Page, Page Template, Save a Page, Find Pages, Items, Route, Approval, Website Editor, and Publish Site without duplicating content or confusing HTML generation with CMS storage.
Where: Flowgen canvas -> Blueprints or Add -> Web & Apps / Flow control
Before you start
- A Website with a connected site identity and collections for CMS steps.
- A designed HTML page for Page Template.
- A list of source records or URLs for batch ingestion.
Key ideas
- HTML and content are separate lanes: Page Template fills a visual page. Save a Page writes a structured entry. A complete workflow may use either one or both, but they are not interchangeable.
- Items owns repetition: Read, fill, save, and query nodes intentionally work on one logical record per run. Items/List iteration supplies typed columns, per-row execution, results, nesting, and history.
- Identity before publication: Find Pages can guard a source URL; Save a Page upserts by slug or fingerprint; Approval can stop a live write; Publish Site delivers the saved Website only after content and template changes are ready.
- Blueprints are editable starting flows: A content-machine blueprint can arrive as one guided card, then expand into the same normal nodes and connections. It is not a separate hidden runtime.
Steps
- Create an Items table with one row per source record or URL and typed columns for everything that varies.
- Inside List iteration, connect the URL to Read a Page and parse/transform the returned Markdown/data.
- Use Find Pages in duplicate-guard mode when the source might already exist. Route new, changed, skip, and review cases explicitly.
- Feed structured values to Save a Page for durable content and, when finished HTML is needed, to Page Template for the visual page.
- Collect results in the Results List and review failed/conflicted rows.
- Place Pause for Approval before any Published/Live content decision.
- Save the Website checkpoint, then run Publish Site to Preview. Verify it; publish Live only on an explicit request.
Recommended graph shape
- Items -> List iteration -> Read a Page.
- Read result -> Transform/research/writer.
- Source URL -> Find Pages duplicate guard -> Route.
- Mapped fields -> Save a Page.
- Mapped fields + designed HTML/CSS -> Page Template.
- Iteration outputs -> Results List.
- Approval -> Publish Site Preview/Live.
Draft-or-Live and Preview-or-Live
- Save a Page controls entry status.
- Publish Site controls delivery target.
- The guided Wizard can collect both answers as explicit visual choices.
- They are independent: a published entry can exist before the site is promoted, and a Preview site can include draft-aware editorial surfaces.
Reliability rules
- Use stable source URLs or slugs.
- Let specific typed wires win over catch-all JSON.
- Review CAS conflicts rather than overwriting.
- Use current graph/list rows at run time.
- Keep live operations behind explicit approval.
- Use the site-bound server runtime for automations.
Tips
- Start with one row and Preview before scaling to the entire list.
- Keep source URL, slug, and publication status as explicit Items columns.
- Store both entry JSON and final path in the Results List for later reruns.
- Use the blueprint when it matches, then inspect and adapt the normal nodes it creates.
Limitations and important notes
- External page access, content rights, and provider availability still apply.
- CMS writes and live publishing require authentication and permissions.
- A page template does not replace a collection schema or a Website document.
- No node in this workflow should silently fan out without Items/List iteration.
Troubleshooting
Every rerun creates a new article
Make slug or source URL stable and connect it to Save a Page on every iteration.
The HTML changed but the CMS did not
Page Template and Save a Page are separate lanes. Connect/run Save a Page when a durable entry must be updated.
Live changed before review
Default Save a Page to Draft and Publish Site to Preview, then place Pause for Approval before changing either target.
Related guides
- Read a Page node: scrape one website into text and data
- Page Template node: turn a designed page into fillable fields
- Save a Page node: create or update one site entry
- Find Pages node: query a site collection or guard duplicates
- Publish Site node: send a saved Website to Preview or Live
- Repeat an Items List and collect typed results