BioFlow

Append ONE new block to the page DRAFT, preserving every existing block byte-for-byte (the copilot's add-preservation path — no whole-array re-emission). Optionally wire an uploaded file from file_list. Requires expectedUpdatedAt from page_get.

page_add_block

Cancel the page's pending scheduled publish (the one page_schedule_publish created), so nothing goes live when it would have fired. The draft and the live page are left exactly as they are. Returns cancelled: false when nothing was scheduled. Works even when the workspace has dangerous operations for AI apps turned off.

page_cancel_scheduled_publish

Create a NEW bio page in the connected workspace as an UNPUBLISHED draft — the same path as the dashboard's New page dialog (slug derived from the title, empty draft, default route provisioned). It never publishes: the page only goes live via page_publish (a separate, dangerous-gated tool). Returns the page id, slug, and draft.updatedAt so you can immediately chain page_update / page_add_block. Repeating the call creates ANOTHER page — pass idempotencyKey to dedupe retries.

page_create

Workspace analytics for the last 7/30/90 days: profile views, clicks, unique visitors, CTR, top links, top referrers, and tip revenue.

analytics_summary

Get one bio page's draft AND published state summaries (block counts by kind, theme, route, timestamps). The draft is what the editor shows; published is what visitors see. The draft section carries updatedAt (pass it as expectedUpdatedAt to every write tool) and the ordered block list with ids/labels (pass those ids to page_update / page_remove_block / page_reorder_blocks).

page_get

List every bio page in the connected workspace: title, slug, publish status, block count, public route, theme, and last update time. Use it first to find a pageId for page_get.

page_list

List the workspace's captured contacts and leads (newsletter signups, form submissions, gated downloads), newest first, paginated via take/skip.

contacts_list

List the workspace's uploaded files (name, MIME type, public path), newest first. Pass a file's publicPath as fileUrl to page_add_block to wire it into an IMAGE/FILE/CAROUSEL/PRODUCT block.

file_list

Publish the page's current DRAFT to its live public route — the same code path as the editor's Publish button. TWO-STEP: the first call (without confirmToken) publishes NOTHING and returns a preview + confirmToken; call again with that token to commit. DANGEROUS: the commit changes what real visitors see, and is refused (DANGEROUS_OPS_DISABLED) unless the workspace enabled dangerous operations for AI apps in Settings -> Connected AI apps.

page_publish

Delete one or more blocks from the page DRAFT by id, preserving every other block, theme, and metadata. Refuses to remove EVERY block (a full-page clear). Requires expectedUpdatedAt from page_get.

page_remove_block