Changelog
What shipped in RunFlowPilot, newest first. Every entry corresponds to a verified production release — nothing is listed before it goes live.
RunFlowPilot rebrand
The product is now RunFlowPilot across the website and app, with the visible name centralized in one configuration.
New features
- Centralized brand name — one setting renames the product everywhere at once.
Improvements
- Consistent RunFlowPilot naming across header, landing, pricing, auth screens, Copilot, dashboard, and notifications.
- Technical identifiers (MCP server name, workflow placeholders, stored deployment records) intentionally unchanged so existing setups keep working.
Documentation page
A searchable documentation hub covering setup, building, deploying, and troubleshooting — every guide audited against the actual app.
New features
- New /docs page with 11 step-by-step guides and a searchable table of contents.
- Footer Documentation link now opens the guides instead of a dead link.
Landing reliability fix
Fixed a server error on the public landing page for the interactive pricing section.
Fixes
- Pricing cards now render client-side behind a loading skeleton, keeping the static landing shell instantly servable.
Pricing sync between landing and app
Public pricing now matches the in-app plans exactly, both reading from one shared plan configuration.
New features
- Single plan catalog shared by the landing page and the in-app upgrade page — prices and features cannot diverge.
- Landing pricing shows Builder, Pro, and Enterprise cards with the monthly/annual selector.
Improvements
- Removed an unimplemented credit-rollover promise from the upgrade page; credits reset each billing period.
Durable deployments with recovery
Deployments now survive failures after n8n creation: retries resume and link the existing workflow instead of creating duplicates.
New features
- Stable per-draft deployment attempts with visible pipeline stages.
- Link-existing recovery for workflows created just before a save failure.
- Setup warnings separated from runnable status — missing credentials show amber, never green.
Precise deploy error reporting
Every deployment stage reports its exact failure instead of a generic server error.
Improvements
- Stage-tagged errors (build, verify, record-save, draft-link) with the n8n workflow ID whenever one exists.
- Post-creation bookkeeping failures surface the created ID with guidance instead of hiding it.
n8n deployment compatibility
Fixed n8n rejecting workflow creation and made rejections fully transparent.
New features
- Failed creates now show n8n's actual message plus the offending field and node.
- Successful creates are read back from the instance and the confirmed ID is shown.
Fixes
- Removed the read-only `active` flag from the create payload, which n8n's schema rejects.
Reliable node saving and preflight
Node edits persist to the server with confirmation, and preflight validates the saved draft.
New features
- Save Changes confirms persistence (“Changes saved”) with field-level errors and no lost edits on failure.
- Preflight re-reads the persisted draft before validating, so it never checks stale data.
Async workflow generation pipeline
Generation moved off the request path onto a queue: submit returns immediately and the app tracks progress to completion.
New features
- Queue-based generation with stage persistence, resumable polling, and a stuck-job watchdog.
Questions about a release? Contact Support.