Privacy Policy
This policy describes how FeedStreak handles personal information, including Google and YouTube user data. FeedStreak is the application at feedstreak.com. It distributes videos you upload to channels you connect. Signing in uses an email address and a password. FeedStreak does not offer Sign in with Google.
Account information
When you create an account, FeedStreak stores your email address, an optional name, and a password hash. Passwords are hashed with scrypt. The password itself is not stored. A session cookie keeps you signed in, and a cookie named active_workspace_id remembers the workspace you are using. Signing out from the app clears distribution drafts stored in that browser.
Workspace members see brands, videos, channels, and publishing records for workspaces they belong to. They do not see another workspace's channels or tokens.
Content you upload
Videos and images you add to Library are stored in a private Cloudflare R2 bucket. They are not published as public files. Downloads use short-lived signed URLs. Optional Library auto-clean is off unless a workspace turns it on (30, 90, or 180 days after the last publish). Items you delete move to Recently deleted and are removed after 30 days. That schedule applies to your media files. It is not a retention period for Google user data.
Publishing records store the caption, title, privacy choice, schedule, and the destination post id you asked FeedStreak to create. Those records stay with the workspace so Queue and history remain accurate.
Google and YouTube User Data
Google user data in FeedStreak is the data received from Google when you connect a YouTube channel, plus the OAuth tokens that authorize that connection. You start that connection from Channels. It is separate from your FeedStreak login.
A normal connect requests only the publishing scopes below. https://www.googleapis.com/auth/yt-analytics.readonly is requested when you choose Enable analytics, and when you reconnect a channel that already has analytics permission. Reconnecting a channel that has never had analytics permission requests the publishing scopes only. The authorization URL asks for offline access and consent so Google can return a refresh token, and it includes previously granted scopes for this app (include_granted_scopes=true).
Publishing scopes, requested on a normal YouTube connect:
https://www.googleapis.com/auth/youtube.readonly— View the YouTube channel on the Google account you connect. FeedStreak reads the channel id, title, custom URL, and thumbnail. When analytics sync runs, this scope is also used to read view, like, and comment counts for videos FeedStreak uploaded.https://www.googleapis.com/auth/youtube.upload— Upload videos you queue in FeedStreak to that channel. The upload includes the video file and the title, description, tags, category, privacy status, and made-for-kids choice you set in FeedStreak.
Publishing scopes plus analytics, requested when analytics is included:
https://www.googleapis.com/auth/youtube.readonly— View the YouTube channel on the Google account you connect. FeedStreak reads the channel id, title, custom URL, and thumbnail. When analytics sync runs, this scope is also used to read view, like, and comment counts for videos FeedStreak uploaded.https://www.googleapis.com/auth/youtube.upload— Upload videos you queue in FeedStreak to that channel. The upload includes the video file and the title, description, tags, category, privacy status, and made-for-kids choice you set in FeedStreak.https://www.googleapis.com/auth/yt-analytics.readonly— Read YouTube Analytics for a video FeedStreak uploaded. The product asks for estimated minutes watched and average view duration. A normal connect does not request this scope.
FeedStreak does not request your Google account email, name, contacts, Drive files, or Gmail. It does not request broader YouTube scopes such as full account management or monetary analytics.
What Google user data is accessed
FeedStreak accesses the following Google user data, and only when you have connected YouTube:
- OAuth access tokens and refresh tokens for the YouTube connection, the token type, expiry of the access token, and the scopes Google granted.
- YouTube channel identity from the YouTube Data API (
channels.list,part=snippet,mine=true): channel id, title, custom URL, and thumbnail. - Upload and processing data for videos you queue: the video file you already stored in FeedStreak, the title, description, tags, category, privacy status, and made-for-kids flag you set, the YouTube video id, the watch URL, and processing status.
- A resumable upload session URI while an upload is in progress.
- When the connection has both analytics scopes, video statistics from the YouTube Data API (view count, like count, and comment count) and YouTube Analytics figures for that video (estimated minutes watched and average view duration). YouTube does not provide a share count through this integration, and FeedStreak stores shares as empty for YouTube.
How Google user data is used
FeedStreak uses Google user data to provide the YouTube features you can see in the product:
- Show the connected channel's name, handle, and avatar on Channels.
- Upload the videos you queue to that channel, using the metadata you entered, and poll until YouTube finishes processing.
- Refresh the access token so scheduled uploads can run later.
- When you have turned on analytics, show views, likes, comments, estimated minutes watched, and average view duration for videos this workspace published.
- Keep a publishing history (video id, watch URL, and the metadata you submitted) so Queue and retry behavior stay tied to the same upload.
FeedStreak does not sell Google user data. It does not use Google user data for advertising, including personalized, retargeted, or interest-based advertising. It does not use Google user data to determine credit-worthiness or for lending. It does not use Google user data to build advertising profiles or databases for sale.
With whom Google user data is shared, transferred, or disclosed
FeedStreak discloses Google user data only as follows. It does not transfer Google user data to advertisers, data brokers, or information resellers.
- Google, because calling the YouTube Data API and YouTube Analytics API is how publishing and analytics work. Video files you chose to publish are sent to YouTube.
- Supabase Postgres, which stores the workspace database for the live service, including encrypted tokens, channel records, publishing records, and analytics snapshots.
- Vercel, which hosts the FeedStreak application. The application decrypts tokens in memory on that host in order to call Google.
- Members of the same FeedStreak workspace, who can see the channel name, handle, avatar, publishing status, and analytics the product displays. They cannot see the OAuth tokens in the product.
These providers are used for other parts of the service and do not receive Google OAuth tokens, YouTube channel records, or YouTube Analytics:
- Cloudflare R2 stores media you upload. Those files are your content. When you publish to YouTube, the application reads the file from R2 and sends it to YouTube. R2 does not store Google tokens.
- Upstash QStash receives a publish job id and a schedule generation number so an upload can start at the scheduled time. The message does not include tokens, channel contents, or analytics.
- Stripe, when billing is turned on, receives the FeedStreak account email and subscription details. It does not receive YouTube data.
No email delivery provider, product-analytics vendor, or error-tracking vendor is installed in the application. An optional caption assistant, described under AI and machine learning, does not receive Google user data.
Data protection mechanisms for sensitive data
YouTube OAuth access tokens and refresh tokens are sensitive. They are encrypted at rest with AES-256-GCM before they are written to the database. Each ciphertext has a random initialization vector and an authentication tag. Additional authenticated data binds the ciphertext to the credential id, the workspace id, the provider (google), and the payload version, so a copied ciphertext does not decrypt for a different workspace. Encryption keys live in server environment variables (SOCIAL_CREDENTIAL_ENCRYPTION_KEY_*), not in the database. Connecting a channel is refused when that encryption is not configured.
Channel names, avatars, granted scopes, analytics numbers, YouTube video ids, and resumable upload session URIs are stored as ordinary database fields. They are not separately encrypted. The production database contract uses TLS (sslmode=require). The application does not currently reject a database URL that omits TLS.
Account passwords are stored only as scrypt hashes. Scheduled-publish messages do not contain tokens. Diagnostic text for the scheduler redacts bearer tokens. Media download URLs are signed and short-lived. The product has no screen that shows OAuth tokens to workspace members.
Data retention and deletion
When you disconnect a YouTube channel, or ask FeedStreak to delete YouTube data, stored YouTube Authorized Data is deleted as soon as possible and within 7 calendar days. That deletion removes data from FeedStreak only. Videos already published stay on YouTube until you delete them in YouTube.
- While a channel stays connected, the encrypted tokens and channel record remain so scheduled uploads and analytics can run.
- If you disconnect a YouTube channel that has no publishing history, FeedStreak deletes that channel row and its encrypted tokens.
- If you disconnect a YouTube channel that already has publishing history, FeedStreak marks that channel row disabled and cancels queued posts that have not started uploading. Publishing history stays on that row's internal id, so those posts cannot be moved onto a different channel. It deletes the YouTube channel title, handle, avatar, granted scopes, analytics, YouTube video ids, watch URLs, and resumable upload session URIs stored for that channel. It also deletes the raw YouTube channel id. The internal id does not identify the YouTube channel, so FeedStreak keeps a server-keyed HMAC-SHA256 of the channel id, only to recognize a later reconnect of the same channel and refuse a different one. When no separate fingerprint key is configured, that key is derived from the sign-in secret, and rotating the sign-in secret changes the fingerprint until the same channel reconnects. That value is still YouTube-related data. The raw id is written back only when the same channel reconnects. The stored value is not shown as the YouTube channel name. FeedStreak keeps the post copy you wrote (title, description, tags, privacy choice, and schedule) and the queue status.
- Disconnect asks Google to revoke the token at https://oauth2.googleapis.com/revoke, using the refresh token when one is stored, otherwise the access token. If another active YouTube channel, including one in a different workspace, still uses that same Google grant, FeedStreak deletes this channel's copy and does not revoke the grant, so the other channel stays connected. The grant is revoked when the last active channel using it is disconnected. If Google does not accept the revoke, the request stays pending, is retried on a schedule, and is visible on the channel. The stored token is deleted when revoke succeeds or when 7 calendar days pass, whichever comes first. You can also remove access at Google account permissions or Google security settings.
- If you revoke FeedStreak in Google Account settings, the next refresh or daily check that receives Google's
invalid_granterror deletes that channel's stored YouTube data within 7 calendar days. Another channel or workspace is not disconnected by that detection. - Once a day, the maintenance job pages through connected YouTube accounts and refreshes each grant. The check does not log tokens. Only a verified
invalid_grantstarts deletion for that channel, including a channel that is never used again. HTTP 429 stops the run and the next day continues. Server errors, timeouts, and network errors leave the channel connected and are retried later. Another channel is not disconnected by that check. - Every maintenance run also finishes pending YouTube deletion and revocation jobs, nearest deadline first, including jobs left over when an earlier run failed or the queue was larger than one page. A job that is already past its 7-day deadline is finished on that run. Overdue jobs are written to the operator log before they are drained.
- Email tryfeedstreak@gmail.com or use the in-app YouTube data deletion request. FeedStreak deletes the remaining YouTube Authorized Data within 7 calendar days. Deleting data stored by FeedStreak does not delete videos or other data stored by YouTube. To delete a video on YouTube, use YouTube.
- There is no in-product control to delete a FeedStreak account, delete a workspace, or export your data.
Publishing records that are not Google user data, such as the title and schedule you entered in FeedStreak, stay with the workspace history. There is no automatic timer that deletes them.
AI and machine learning
FeedStreak does not use Google user data to develop, improve, or train generalized or non-personalized AI or machine-learning models. Google Workspace APIs are not used. YouTube data is not sent to a model for training.
An optional caption assistant can be turned on with server configuration. It is off unless an API URL, key, and model are all set. When it runs, the request contains the destination platform name, a video title you supplied, and any brand voice or keywords you supplied. It does not contain Google OAuth tokens, YouTube channel ids, channel names, or YouTube Analytics. The reply is a caption suggestion only. It does not publish.
Limited Use
FeedStreak's use of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.
Google user data is used only to provide and improve the user-facing YouTube connection, publishing, and optional analytics features described in this policy. It is transferred only to the processors named above, and only so those features can run. FeedStreak does not sell Google user data and does not use it for advertising. Humans are not given a product feature that displays OAuth tokens. Workspace members can see the channel identity and analytics the product shows for their own workspace. FeedStreak does not allow other human reading of Google user data except with your consent, for security (such as investigating abuse), or when the law requires it. Google user data is not used to train generalized AI or machine-learning models.
Use of YouTube API Services is subject to the YouTube Terms of Service. Google's handling of Google account data is described in the Google Privacy Policy. You can review or remove this app's access at https://myaccount.google.com/permissions.
Other destination accounts
You can also connect other publishing destinations you choose, such as TikTok, Instagram, Facebook, LinkedIn, Threads, Pinterest, and X. Their tokens are stored in the same encrypted credential store. Disconnecting a destination follows the same local rules as YouTube: delete the row when it has no publishing history, or disable it and keep history when it does. YouTube disconnect also revokes the Google token and deletes the stored ciphertext. TikTok and X revoke their own tokens on disconnect. A YouTube channel with publishing history keeps the non-secret row after its tokens are deleted.
Changes
If FeedStreak changes how it uses Google user data, this page will be updated and the date above will change. The current policy is always the page at /privacy.
Contact
Email tryfeedstreak@gmail.com with a privacy question or a request to delete YouTube Authorized Data stored by FeedStreak. That deletion is completed within 7 calendar days. It does not delete videos that remain on YouTube. Recovery steps are on Support.