Data and access

Privacy policy

Last updated:

Alexey Dubkov Publisher is operated for Alexey Dubkov’s personal YouTube channel by the channel owner and authorized project operator. It is an owner-operated tool, not an application open to public registration. This policy covers video transfer, verification and the optional AI analysis and metadata editor described below. Each optional feature requires separate enablement; this policy does not state that it is currently active.

1. Data accessed

Telegram

The tool checks messages in one configured private chat and selects only incoming video messages from the configured sender. For qualifying videos, it uses the chat, message and media identifiers, file type, file size and the video file. These identifiers allow interrupted transfers to resume and help prevent duplicates.

Telegram authorizes a user-client session at the account level; it does not give this application a separate permission limited to one chat. The application enforces its single-source restriction in its own code. Source message captions are not used to generate titles or descriptions and are not copied into the transfer queue.

Google and YouTube

  • youtube.upload: uploads the authorized video and its generated title, description, tags, category and selected visibility setting.
  • youtube.readonly: reads the authorized channel’s identity and the uploaded video’s identifier, channel, processing and privacy status to confirm the destination and result.
  • youtube.force-ssl, only for the separately authorized metadata editor: reads existing video metadata and updates titles, descriptions and tags on the configured channel. Although Google describes a broader permission, this editor is limited to snippet metadata updates; it does not upload or delete videos or modify their visibility or status.

Google OAuth access and refresh tokens are stored so the authorized worker can make these requests and continue operating without repeating consent for every transfer. The application uses Google’s authorization flow and does not collect the Google Account password.

2. Purpose, use and transfer

Data is used to discover qualifying videos, download temporary copies, perform technical file checks, identify duplicates, prepare metadata, upload to the configured YouTube channel and confirm successful processing and the intended visibility. Without optional AI analysis, titles, descriptions and tags use neutral templates.

The video file and generated metadata are sent to YouTube through its API. When the owner selects public visibility, the video and its title, description and tags become publicly available on the channel. The owner has authorized public archive uploads in the current workflow. Telegram processes the requests needed to retrieve the source material. The project’s VPS processes the temporary video and operational records. Google, YouTube and Telegram handle data under their own applicable policies, including Google’s Privacy Policy and Telegram’s Privacy Policy.

Project access is limited to the owner and authorized operator for transfer, support and maintenance. The application does not sell Google user data, use it for advertising, or use it to train general-purpose AI models.

Use of Google API data is limited to the workflow described here and follows the Google API Services User Data Policy, including its Limited Use requirements.

Optional Gemini video analysis

When enabled, a temporary copy of the original authorized Telegram video, including its audio, is uploaded to Google’s paid Gemini API. The API receives the video and instructions to draft a Russian title, a one- or two-sentence description, three to five hashtags and brief supporting observations about visible action or clearly audible speech. Source message captions are not supplied for this analysis.

The analysis is instructed not to identify people, infer identities or include names or geographic locations. Output checks reject unclear or unsuitable responses; if analysis is unavailable, the transfer can continue using neutral metadata. These controls do not guarantee that generated text is error-free. Metadata may be applied automatically to new uploads or, when the separate editor is enabled, to existing videos on the same channel. Applied metadata is public when the video is public.

Under Google’s paid Gemini API terms, submitted content and responses are not used to improve Google’s products. Google may retain prompts and responses for a limited period for safety, abuse prevention and required legal disclosures. This is separate from the temporary uploaded-file lifetime described below.

3. Storage, protection and retention

The worker runs under a separate Linux service user in the project’s VPS directories. Credential and session files use owner-only file permissions (0600), and access is restricted through filesystem controls. Credentials are not part of this website or the public application files. Tokens and resumable-upload session URLs are excluded from normal application logs.

  • Temporary video copies: removed after YouTube confirms the expected channel, completed processing and the visibility recorded for that transfer. A copy may remain while transfer, verification or a failure review is unresolved.
  • Gemini uploaded files, when analysis is enabled: the client attempts to delete only the file it uploaded when analysis ends, including after an error. An interrupted process, failed deletion or lost upload-finalization response can prevent immediate deletion or confirmation; if the returned file identifier is unknown, the client does not guess or delete other files. Google’s Files API documentation states that uploaded files are automatically deleted after 48 hours. This file expiry does not describe separate request or security logs.
  • Operational records: source identifiers, SHA-256 file hashes, file sizes, technical video details, generated metadata, YouTube video identifiers, upload checkpoints and job status are retained to recover work and prevent duplicate uploads. When analysis is enabled, its cache also retains the SHA, model and prompt version, resulting text and observations, and budget reservations or recorded usage. These records avoid repeated paid analysis and preserve the spending limit across restarts. There is currently no automatic time-based deletion of these records.
  • Credentials: local credential and session files are retained until the operator removes them. Revoking Google access or terminating the Telegram session stops that access but does not automatically erase these local files.

The application does not delete original Telegram messages or source videos. Removing a temporary copy or queue record does not remove a video already uploaded to YouTube.

4. Stop access or request deletion

  • Ask the operator to pause transfers, remove project-held temporary files or review deletion of retained operational records using the contact address below.
  • Revoke Google access through your Google Account’s third-party connections. Revocation stops further authorized API access; it does not automatically erase the project’s existing records.
  • Terminate the application’s separate Telegram session from Telegram’s Settings → Devices or Active Sessions.
  • Manage or delete uploaded videos using the owner’s YouTube account. Revoking this application’s access does not delete those uploads.
  • Ask the operator to disable optional AI analysis or metadata editing independently. The owner or operator can also revoke the relevant Gemini API key or the editor’s separate Google authorization. This does not automatically remove cached results or metadata already applied to a video.

Removing queue records may prevent the tool from recognizing previously transferred files. The owner and operator can review that consequence before resuming work.

5. These information pages

These static pages contain no registration form, analytics scripts, advertising, tracking pixels or embedded third-party media. The hosting server processes IP addresses, request paths and browser headers to deliver the pages. Routine access logging is disabled for this site. Technical error logs may contain request details used to diagnose failures and are retained for maintenance until rotated or removed by the operator.

If you contact the support email, the project uses the information you provide to handle your request; the email provider also processes that correspondence.

6. Contact and changes

For support, access questions or data-removal requests, email .

This page will be updated if the workflow or data practices change. New uses of Google user data will be disclosed before they begin, with renewed authorization where required.