Know what PageFluxer captures, where it is kept and who can see it. Here is how monitoring, retention and your choices fit together.
Monitored content can contain information from the source page. Choose public pages and selections you are authorized to monitor, and consider what will be visible to your team and notification destinations.
The application database stores monitored URLs, labels, selectors, rules, recorded public steps, successful text captures and retained changes. PDF monitoring stores extracted readable text for comparison. Saved AI summaries follow their change history.
Visual baselines, change images and pixel diffs use private object storage. Postgres stores references and comparison metadata. Viewing an image requires workspace access before a short-lived image link is issued.
Public URLs and capture settings are processed by the app or isolated browser service. Browser sessions are temporary. The Chrome extension passes an explicit monitoring draft; it does not transfer your login session or cookies.
Recorded steps can contain public search text and chosen values. Those values may also appear in a capture. Use them for public page interactions; do not enter passwords, payment information or other confidential content. Recorded steps guide →
History access follows the owning workspace’s plan. Older history may be removed earlier when a shared storage allowance is reached. Pausing a monitor does not stop retention cleanup.
| Plan | Text history | Text storage | Image history | Image storage |
|---|---|---|---|---|
| Free | Up to 7 days | 20 MiB | Up to 7 days | 50 MiB |
| Starter | Up to 30 days | 100 MiB | Up to 14 days | 250 MiB |
| Pro | Up to 90 days | 500 MiB | Up to 30 days | 1 GiB |
| Business | Up to 180 days | 1 GiB | Up to 30 days | 3 GiB |
Baselines are different from history. The latest successful text baseline remains while the monitor exists so the next check has a comparison point. Active visual baselines are renewed; an unrefreshed visual baseline expires after 90 days.
Previews expire sooner. Stored setup preview HTML and visual-preview access expire after 15 minutes. Background cleanup removes expired records and image objects afterward.
Changing plans affects access. A downgrade applies the shorter history window immediately. Upgrading cannot restore captures that cleanup has already deleted. Visual access also depends on feature and trial allowances.
Service records have other periods. Completed notification delivery logs are kept for seven days; bounded readiness totals for 90 days; monthly usage totals for up to 13 calendar months. The privacy policy covers account, session and other operational records.
Deleting a monitor removes its captures and change history from the active application database. Its screenshot objects are queued for deletion; background cleanup and configured storage lifecycle rules remove them afterward.
Private image links already issued can remain usable for up to 60 seconds after an access change. Do not share them as permanent links.
For account deletion, correction or access requests, contact support@pagefluxer.com. We verify account control and, for team data, the requester’s authority. Account deletion is handled through support.
When you connect an MCP assistant, you choose a workspace and approve its access. Granted monitor settings, captured text and saved summaries can be sent to that assistant provider. Its own privacy and retention policies apply to those copies. Connections use scoped OAuth tokens and expire after 30 days; you can revoke them sooner in Assistant connections. Removing or changing a member’s role revokes their connections.
PageFluxer stores token hashes, connection permissions, short-lived retry receipts, and aggregate operation counts. It does not send captured text or credentials to analytics. Connecting an assistant does not enable automatic OpenAI summaries or transfer browser logins and cookies. Assistant permission guide →
Regular capture, comparison and notification rules do not require OpenAI. Optional AI setup and AI summaries have separate consent controls.
After you consent to a request, your prompt, selected captured text, candidate labels and supported capabilities are sent to OpenAI for an editable suggestion. This does not enable automatic change summaries.
AI setup details →Automatic summaries are off by default. Enable them with consent per monitor, or give permission for a single saved change. Selected before/after text and relevant change evidence are sent; saved summaries may appear in notifications.
Summary consent and controls →These requests do not send screenshots or browser cookies as input fields. Common sensitive patterns in excerpts are filtered, but filtering cannot guarantee removal of confidential information. Entered text may appear in a captured excerpt.
Provider data handling →The API requests use store: false to disable response storage as application state. That setting is not a promise of zero provider retention. Turning summaries off stops new automatic requests and cancels queued work; it cannot recall an in-flight request or erase already delivered messages. Existing summaries remain subject to normal history deletion.
Joining a Business team does not expose your personal workspace to that team. Within the shared workspace, members can see all monitors, captured content, retained history and review state, including reviewer identities. There are no private per-monitor permissions within that workspace.
| Action | Owner | Admin | Member |
|---|---|---|---|
| View shared captures and history | Yes | Yes | Yes |
| Create, edit and delete monitors; review changes | Yes | Yes | Yes |
| Connect destinations and manage API tokens | Yes | Yes | No |
| Invite and manage ordinary members | Yes | Yes | No |
| Assign or change administrator roles | Yes | No | No |
| Manage billing | Yes | No | No |
Removing a teammate prevents new workspace access but preserves the shared monitors and reviews. Notification channels can include people outside PageFluxer; everyone who can read the destination may see delivered content. Workspace API tokens grant their configured scopes until revoked or expired; review tokens separately when access needs change.
Workspace roles and invitations →Sign-in uses short-lived email links. Session cookies are HTTP-only, use SameSite protections and are marked Secure over HTTPS. Login and API tokens are stored as hashes rather than reusable plaintext tokens.
Stored webhook signing secrets and team-channel connection URLs are encrypted. API tokens have scopes and expiry controls; signed webhooks let receivers verify the payload. Protect your email account and the tools that receive your notifications.
API authentication and permissions →Railway hosts the application and Postgres database. Browser checks use an isolated renderer; visual monitoring uses private S3-compatible object storage. Resend delivers email and Stripe handles payment details. PageFluxer keeps billing identifiers and limited billing location fields, not your full card details.
Optional OpenAI features process consented inputs. Connected Slack or Microsoft Teams destinations receive the notifications you enable. Sentry processes filtered diagnostics where configured. Google processes optional analytics when accepted, and the website loads Google Fonts.
Processing can take place outside Canada. Provider regions, backup schedules and provider-specific log retention depend on the deployed services. Contact us for current details before using PageFluxer for data with residency or special handling requirements.
Full privacy policy and provider details →Ask about storage, access, deletion or a security concern. For a suspected vulnerability, send a brief description and a safe way to reproduce it; leave credentials and sensitive captures out of the first message.