Question 1

What does Live data retain?

Accepted Answer

New projects default to Live data. Datapipe fetches and processes report rows to show a result or send it to your chosen destination, without saving those rows or run previews in Datapipe. We still retain account and billing details, encrypted connection credentials, report definitions, selected resource IDs, schedules, and operational metadata such as status, row counts, and timestamps. Live means fetched when requested or refreshed, not a guaranteed real-time stream. Source availability, quotas, and historical limits apply.

Question 2

How do I choose Saved history?

Accepted Answer

Select Save history when creating a project, or open that project's Settings and enable Saved history. Scheduled report refreshes can then retain rows: 90 days on Free, or while subscribed on paid plans. Stored run previews are also enabled. CSV uploads and captured webhook events require Saved history because their inputs must be stored; they remain until removed, switching to Live data, or account deletion. Existing projects keep their previous Saved history setting. Enabling history does not recover data that the source no longer provides, and not every preview is archived. Data Vault is the optional $9/month backup and extended-retention add-on; it never turns history on automatically.

Question 3

How are connection credentials protected?

Accepted Answer

The production website uses HTTPS. OAuth tokens and pasted API credentials are encrypted with AES-256-GCM before being saved. Datapipe decrypts them on the server to make authorized API requests; this is not end-to-end encryption. OAuth sign-in happens with the provider, so Datapipe does not need that provider's password. For a key fallback, create a dedicated restricted key with the minimum required permissions. Account passwords are hashed using scrypt. Protect your provider account with multi-factor authentication and never include keys or tokens in a support ticket.

Question 4

What permissions do connections need?

Accepted Answer

Google account sign-in verifies identity separately from reporting access. Google source authorization requests the selected product's permissions; YouTube uses a separate grant for channels and Brand Accounts. Reporting connectors read source data. Destinations such as Sheets and Drive need permission to create or update the reports you request. Some providers offer broader permission scopes than a single report needs. Review the consent screen. Custom integrations use the endpoint, method, and credentials you supply; a POST endpoint may change provider data, so choose an appropriate reporting endpoint and restricted credentials.

Question 5

Can someone see another project's data?

Accepted Answer

Connections, reports, history, and AI tokens are scoped to a selected project. Server-side access checks verify workspace and project ownership; changing a URL does not grant permission to another workspace. Invited read-only logins are limited to their project. Keep invitation links private. Datapipe's admin tools require a verified owner session. These are application access controls, not a separate physical database for every project.

Question 6

How do disconnecting and deleting differ?

Accepted Answer

Disconnecting a connection stops future source reads and releases its connection slot. It does not erase previously saved history or copies already delivered elsewhere. Where multiple Google connections share one authorization, the grant may remain for those connections; revoke Datapipe in the provider's account settings to remove provider-wide permission. To remove a project's saved rows, open Settings, choose Switch to Live data, and confirm deletion. This removes managed history, run previews, CSV uploads, and captured webhook events from active storage while keeping report definitions and connections. A failed deletion is shown as pending and can be retried. Infrastructure backups follow the host's retention, so deletion does not promise immediate physical erasure. Full account deletion is available in Settings or through a privacy request.

Question 7

What happens when I send data to Sheets, BigQuery, Claude, or ChatGPT?

Accepted Answer

A destination receives the data you choose to send, and can retain its own copy even in Live data mode. Looker Studio uses a Google Sheet. Your own BigQuery destination is on Pro and Agency. Claude and ChatGPT access uses read-only credentials scoped to one project; switching the app's project does not broaden that token. Enabling Saved history is separate from authorizing any destination. AI access can still reveal sensitive results to the assistant you connect. Review the destination's permissions, retention, and sharing settings. Deleting data in Datapipe does not delete destination files, downloads, or AI conversations. Datapipe does not use connected Google user data to train generalized AI models.

Question 8

Where is data processed, and what security claims can Datapipe make?

Accepted Answer

The app is hosted on Fly.io in the United States. Optional saved history uses Datapipe's BigQuery when configured, with a local SQLite fallback. Connection credentials have application-level encryption; saved report rows do not have the same application-level field encryption. We do not describe the service as end-to-end encrypted or claim a SOC 2 or ISO 27001 certification. Security controls reduce risk and cannot guarantee zero incidents. Our privacy policy and subprocessor list describe service providers and processing. For a security concern or privacy request, use Contact us and choose the appropriate category, or email info@trydatapipe.com. Share a description, not secrets or customer datasets.

Connect and report by default. Retain history by choice.

Live data by default

Saved history when you choose