In short
Raw writing samples used to build a profile are used once and discarded. Source text you paste into Generate isn't saved just by generating — but if you save an output, its supporting source and review evidence are stored with it, where available. What else persists is the profiles you build and the outputs you explicitly choose to save, held in a workspace only you can access. Operational logs record metadata, never your writing.
Content Flow is a private tool for creating reusable writing style profiles and drafting content in those styles. This page explains, in plain English, what happens to the material you put into the tool and what actually gets stored.
Writing samples
When you create a style profile, you paste authorised writing samples that represent the communication style you want the profile to learn from. Those samples are sent once to the analysis service, which reads them for tone, structure and communication patterns. The samples themselves are then discarded — not stored, not logged, not written to the Content Flow database.
What's kept is the profile the analysis produces: the reusable style guidance, not the source writing. The samples used to build it do not persist.
Source text you paste into Generate
When you generate a draft, you paste source text — the facts, points or brief you want the draft to cover. That text is sent to the generation and review services only when you trigger those actions. Generating alone doesn't save it — it isn't written to the database or logged. If you then save the output, though, the source text and the source details identified in it are stored with that saved output — see Generated drafts below.
Worth knowing
Until you save the output, if you refresh the page or navigate away, the source text is gone. You'd need to paste it again next time — while it's unsaved, it isn't being held anywhere.
Generated drafts
The draft Content Flow produces is session-only by default. Close the tab without saving and it disappears.
If you click Save output, the draft is stored in your signed-in workspace, along with which profile was used and how the review checks came out. For evidence-backed outputs, the source text you pasted is stored with the saved output too, together with the Source Detail Ledger and its exact source quotes, the output's lineage, and the available Review evidence. Some outputs — such as those saved from a multiple-output pack — may not carry this supporting evidence.
What is stored in your signed-in workspace
When you sign in, Content Flow connects to a workspace that only you can read from and write to. That workspace holds:
- Style profiles you've created — the Personal Voice or Brand Voice profiles describing the writing style for a person or brand.
- Saved outputs you've explicitly chosen to save.
- Versions — if you edit a saved output as a new version, the edit is kept as a linked version of the original.
- Projects — optional groupings you can use to organise saved outputs.
- Supporting evidence for saved outputs — where available, the source text you pasted, the source details identified in it and their exact quotes, the output's lineage, and the Review result — stored with the saved output it belongs to.
Each of these is scoped to your account. The database is set up so signed-in users can only access the workspace data they're allowed to — enforced at the database level, not just hidden in the app. The API key that lets Content Flow generate content lives only on the server side; it's never sent to your browser or exposed in any page.
Deletion and archive
Normal app controls archive items rather than deleting them. Archiving is not deletion: archived outputs and their supporting evidence remain stored and can be restored — for saved outputs, through the Archived view. Archived items are hidden from everyday views, not removed.
Archived saved outputs can be permanently deleted from the Archived view. Permanent deletion removes the selected output, its saved versions and associated unshared supporting evidence from Content Flow's live application tables, and the output cannot then be restored. Content held temporarily in Supabase backups or point-in-time recovery, or retained by an external service under its own retention arrangements, is not necessarily erased at the same moment. Nothing is deleted automatically. For other removal requests, use the contact page; requests are assessed and handled in accordance with applicable data-protection rights.
What Content Flow doesn't do
- No analytics tracking on your usage.
- No third-party cookies.
- No pixel trackers or session replay.
- No sharing of your data beyond the analysis, generation and review services — which are called only when you trigger them.
- No advertising anywhere in the tool.
Your part
Only paste writing samples or source material you're authorised to use. Content Flow's storage rules limit what's kept, but authorisation to use the material in the first place is always yours to confirm — not something the tool can check for you.