Project Radar Privacy Policy
Summary
This Policy applies to Project Radar, the web application at project-radar.ru: its public pages and the app in which team members work.
- Project Radar collects comments and direct messages addressed to a team's own connected social media pages into one feed, where the team replies to and moderates them.
- Project Radar uses YouTube API Services. From YouTube, Project Radar reads the connected channel's basic details, the details of its videos and the comments on them. It replies to or moderates comments only when a team member acts, under a rule set up by the workspace owner or, if the owner turns it on, in automatic spam mode (section 9).
- Project Radar refreshes YouTube data at least every 30 days and deletes it no later than 7 days after the channel is disconnected or access is revoked (section 9.5).
- Data is stored on a server in Russia. If AI features are enabled (they are off by default), the text of incoming messages is sent to Anthropic, which does not use it to train models (section 5).
- No ads, no selling of data, no data brokers, no analytics trackers. Project Radar itself sets no cookies. Project Radar does not train AI models and does not generate images.
- To delete your data or ask a question, write to privacy@gomon.space (section 13).
1. Who processes the data
Project Radar (project-radar.ru) is a tool for teams that run project pages on social networks: it collects comments and direct messages into one feed, where team members reply to and moderate them.
This Policy applies to the Project Radar web application at https://project-radar.ru, including its public pages and the app used by team members, and to the data Project Radar receives through the APIs of the social networks connected to it.
The operator of the service is Individual Entrepreneur Lev Aleksandrovich Evseev (INN 540440564800), hereinafter “we”. For personal data matters, contact us at privacy@gomon.space.
This Policy has been prepared in accordance with Federal Law No. 152-FZ of 27.07.2006 “On Personal Data” and describes data processing in the Project Radar service. By using the service, you confirm that you have read it.
Project Radar is organized into workspaces: a workspace is a project or company with its own social media pages and its own team members. We process the audience data of a page connected by the workspace owner on the owner's instructions and only to operate the service.
2. What data we process
2.1. Team members who have a Project Radar account
- name, email address, role and the workspaces they have access to;
- password — only as an irreversible hash; we do not store or see the password itself;
- Telegram ID and avatar, if the team member has enabled notifications in Telegram;
- sign-in log: time, IP address, browser and device information;
- log of actions in the service: replies (with the first 200 characters of the text), status changes, deletions, bans, settings changes;
- interface settings.
2.2. Data from connected social media pages
- posts of the connected page;
- comments on these posts and messages that social network users themselves sent to this page or to chats in which its bot is a member;
- public author profile data provided by the platform: name, username, identifier, profile link, avatar, and sometimes the number of followers;
- attachments to comments and messages: images, videos, documents;
- copies of authors' avatar images saved on Project Radar's server;
- team members' replies and the processing history of each incoming message;
- access tokens for connected pages — stored on the server and never shown in the interface.
Project Radar does not access social network users' private conversations with each other, the non-public data of their profiles, or pages that are not connected to a workspace.
2.3. Technical data
- IP address, request time and browser information in server logs;
- Project Radar itself does not set cookies; the browser stores only the session token and interface settings (localStorage);
- no analytics services or advertising counters are installed on the site;
- the pages load the Manrope font from Google Fonts (Google LLC), which receives the visitor's IP address and browser information;
- inside the app, video thumbnails and some attachments are loaded directly from the platforms' servers, and a video attached to a message may be shown in the platform's own embedded player (YouTube, VK, TikTok). The player is loaded from that platform and may store its own cookies under the platform's privacy policy; Project Radar does not read them.
3. Why we process data
- to show workspace team members the comments and messages addressed to their page and let them reply to, hide or delete them;
- to protect pages from spam: block an author and delete their messages upon a team member's action, under a workspace rule or in the automatic spam mode the workspace owner can turn on (section 5);
- to keep a shared ban list of Telegram spammers: the operator can ban a spammer's Telegram account in the chats of all workspaces that take part in the list. Only Telegram authors can be put on this list; YouTube data is never used for it;
- to calculate team performance statistics: response speed, workload, sentiment of incoming messages;
- to send team members notifications about new incoming messages and failures;
- to ensure security: protection against password guessing, sign-in and action logs, incident investigation;
- to respond to user requests and comply with legal requirements.
We do not sell data, do not use it for advertising, do not share it with ad networks or data brokers, and do not train artificial intelligence models on it.
4. Legal grounds
- for team members — the Terms of Service and the instruction of the workspace owner who creates accounts for their team (Part 3 of Article 6 of 152-FZ);
- for authors of comments and messages — the instruction of the workspace owner who connected their page (Part 3 of Article 6 of 152-FZ), as well as the legitimate interest of the page owner in responding to messages addressed to them and moderating their page (Clause 7 of Part 1 of Article 6 of 152-FZ);
- for security logs — the operator's legitimate interest in protecting the service.
The workspace owner confirms that they are entitled to manage the page being connected and to instruct us to process the data of its audience.
5. Automated labeling (AI)
AI features are off by default. If the workspace owner has enabled them, the text of a new comment or message is sent to a language model to determine its topic, sentiment and signs of spam, and to recommend an action (close without reply, reply needed, delete or escalate) with a short reason. If image checking is enabled (Telegram only), the image from the incoming message is sent as well.
- The model is Claude by Anthropic PBC (USA), accessed through its API. Under the terms of its commercial API, Anthropic does not use this data to train its models. If Claude is unavailable or its daily budget is used up, an open model running on Project Radar's own server may be used instead; in that case the text does not leave the server.
- Only the content of the incoming message and the minimum necessary context are sent to the model. Access tokens, passwords and team members' email addresses are not sent.
- By default the labeling result is only a recommendation. Decisions to reply, delete or ban are made by a team member or by a rule set up by the workspace owner.
- Automatic mode. The workspace owner can switch spam handling to automatic mode: then Project Radar itself hides (holds for review) a comment it is confident is spam, within an hourly limit the owner sets. In Telegram chats the owner can also allow Project Radar to delete such a message and ban its author in that chat. The owner can likewise let Project Radar close messages the model is confident need no reply; this only changes their status in Project Radar. Every such action is logged; hidden comments can be shown again and closed messages reopened. AI never publishes replies and never bans YouTube commenters.
- The workspace owner can turn AI off completely, in which case incoming messages are not sent anywhere.
- Help assistant. When a team member asks the built-in Help assistant a question, the question and the recent conversation with the assistant are sent to Claude together with the help articles. No incoming messages, Google user data or other workspace data are included. The operator turns the assistant on or off for the whole service.
- Project Radar itself does not train any AI models and does not generate replies, images, video or audio.
6. Who receives the data
- Social networks (Telegram, VK (VKontakte), Google/YouTube, Meta/Instagram, TikTok) — when a team member replies to, hides or deletes a comment or blocks an author. The data is sent through the official API of the platform where the incoming message was published.
- Anthropic PBC (USA) — for automated labeling, if it is enabled, and for the Help assistant (section 5). This is a cross-border transfer, and it is limited to the content of incoming messages and the questions team members ask the Help assistant.
- Hosting. Project Radar's servers and database are located in Russia. For stable access from abroad and via VPN, a network node in Sweden is used: it relays encrypted traffic between the browser and the server without decrypting it. Requests to the APIs of foreign platforms may also go through an encrypted channel abroad.
- Google Fonts and platform content. The browser loads the font from Google Fonts and, inside the app, video thumbnails, some attachments and embedded players directly from the platforms; they receive the IP address and browser information (section 2.3).
- Email addresses set by the workspace. If a workspace administrator has turned on email forwarding (Settings → Email forwarding), a team member can click “Forward” to send a comment or conversation (text, author name, link) by email to these addresses. Project Radar forwards only when a team member clicks; the emails stay with the recipients.
- Government authorities — only upon a lawful request and to the extent established by law.
There are no other recipients. Team members of one workspace cannot see the data of other workspaces. The operator's service administrator account has technical access to every workspace for support and security; it opens a workspace's data only in the cases listed in section 9.7.
7. Where and how long we store data
- The database is stored on a server in Russia.
- Comments and messages are stored while the workspace uses the service and the page remains connected. If a connected page is deleted together with its data, its comments and messages are deleted. If it is disconnected with its history kept, the history remains in the workspace until the owner deletes it.
- Data received from YouTube has shorter limits: Project Radar refreshes it at least every 30 days, deletes anything it could not refresh, and deletes everything received from YouTube no later than 7 days after the channel is disconnected or access is revoked, including when it is disconnected with its history kept. Workspace statistics are calculated from stored comments, so deleted comments disappear from them as well (section 9.5).
- YouTube author profile records are deleted automatically under the rules of section 9.5. Author records from other social networks (name, username, ID, profile link, avatar link) are kept in a shared table of authors and are not removed automatically when a page or workspace is deleted; we delete them when we handle a deletion request (section 13). The saved copy of an author's avatar is deleted automatically once Project Radar no longer stores any comments or messages from that author.
- A team member's account is stored while they work in the workspace. At the request of the workspace owner or the team member, it is deleted or blocked.
- The sign-in log is stored for 90 days.
- The action log (who did what and when, including the first 200 characters of each reply sent) is kept while the workspace uses Project Radar and is deleted together with the workspace. It is not removed when a single page is deleted.
- Automatic database backups: daily backups are kept for 7 days, weekly backups for 4 weeks. After that they are overwritten, so deleted data disappears from them no later than 5 weeks after deletion. Before maintenance the operator can also make a manual backup: the 10 most recent manual backups are kept without a time limit, and older ones are deleted. If recent backups fail the integrity check, the last intact backup is kept until a newer intact one appears.
- If the law requires data to be stored longer, we store it for as long as the law requires.
8. How we protect data
- all connections to the service are encrypted (HTTPS);
- passwords are stored only as bcrypt hashes; a password issued by an administrator must be changed at first sign-in;
- permissions are separated by role and workspace, and every action is recorded in a log;
- social network tokens are stored only on the server and never appear in the interface or logs;
- only the operator has access to the server; the service runs with minimal system privileges.
9. Google user data (YouTube)
Project Radar uses YouTube API Services. By connecting a YouTube channel to Project Radar, you agree to be bound by the YouTube Terms of Service. Google's own processing of your data is governed by the Google Privacy Policy.
This section explains how Project Radar accesses, uses, stores, shares and deletes Google user data received through the YouTube Data API. It applies when the owner or an administrator of a Project Radar workspace connects their own YouTube channel with “Sign in with Google”.
Project Radar requests a single OAuth scope, https://www.googleapis.com/auth/youtube.force-ssl, with offline access, so that new comments keep arriving without signing in again. Project Radar does not request any other Google scopes, so it does not receive the Google account's email address or data from other Google services. Project Radar uses this scope only for the comment features described in this section.
9.1. Data we access
- Connected channel (
channels.listwithmine=true, partssnippetandstatistics): channel ID, title, handle (custom URL), channel link, avatar and subscriber count. The same request is used to refresh this data (section 9.5), and achannels.listrequest with partidis used to check that the connection is working. - Videos (
videos.list, partsnippet): title, thumbnail link, publish date and link of the channel's videos that comments were posted under. Avideos.listrequest by video ID is also used to refresh the stored video details (section 9.5). - Comments and replies on the channel's videos (
commentThreads.list, partssnippetandreplies;comments.list): comment ID, comment text, author display name, author channel ID and channel URL, author avatar URL, like count, which comment it replies to, time of publication and of the last update. Acomments.listrequest by comment ID is also used to refresh stored comments, andcommentThreads.listwithmoderationStatus=heldForReviewto refresh comments that team members held for review (section 9.5). - Author avatars. Project Radar saves a copy of each comment author's YouTube avatar image (a 176-pixel version from the same link) on its server, so the feed does not depend on expiring links. The copy is downloaded again when the avatar link changes (at most once a week) and in any case at least every 30 days; a copy that could not be refreshed within 30 days is deleted.
- OAuth tokens that Google issues for the channel (access token and refresh token).
Project Radar does not upload or edit videos, does not change ratings or captions, does not access private messages and does not receive data of channels that are not connected.
9.2. How we use it
- Channel ID, title, handle, link, avatar and subscriber count — to show the workspace which channel is connected and whether the connection works (the account card in Settings → Accounts).
- Video details — to show next to each comment which video it belongs to, with its thumbnail and a link to it.
- Comments, replies and author data — Project Radar checks the channel for new comments periodically and shows them in the workspace's feed, where team members read and answer them. The same records are used for the workspace's reply and moderation rules, its statistics (response speed, workload, sentiment) and, if enabled, AI labeling (section 5).
- Avatar copies — to show commenters' avatars in the feed.
- OAuth tokens — only to call the YouTube Data API on behalf of the connected channel for the purposes in this list.
Project Radar performs actions on the channel only when a team member clicks the corresponding button, under a reply or moderation rule set up by the workspace owner or, if the owner has turned on automatic spam mode, to hide a comment it is confident is spam (section 5):
comments.insert— publish a reply written by a team member on behalf of the channel;comments.setModerationStatus(heldForReview,publishedorrejected) — hold a comment for review, publish it again or reject it; withbanAuthor=trueon the author's latest comment, also ban the commenter from the channel;comments.delete— delete a reply that the connected channel itself posted; other people's comments are removed withsetModerationStatus=rejected.
9.3. Storage and protection
- Google user data is stored in Project Radar's database on a server in Russia. Copies of commenters' avatar images are stored as files on the same server.
- OAuth tokens are stored on the server only. They are never shown in the interface and are never written to logs.
- Only team members of the workspace that connected the channel can see its data, within their roles. Team members of other workspaces cannot see it. The operator's service administrator account has technical access to every workspace for support and security; it opens a workspace's data only in the cases listed in section 9.7.
- All connections are encrypted (HTTPS). Other security measures are described in section 8.
9.4. Sharing and disclosure
We share Google user data only in these cases:
- YouTube (Google). Replies and moderation actions are sent to the YouTube Data API on behalf of the channel when a team member or the owner's rule triggers them.
- Anthropic PBC (USA), only if the workspace owner has AI features enabled. The text of a new comment and the minimum context needed are sent to the Claude API to determine its topic, sentiment and signs of spam and to recommend an action. OAuth tokens, passwords and team members' email addresses are never sent. Under its commercial API terms, Anthropic does not use this data to train its models. If AI is turned off, nothing is sent. If Claude is unavailable, the comment may be processed by an open model on Project Radar's own server, and nothing leaves the server.
- Email addresses set by the workspace. If a workspace administrator has turned on email forwarding, a team member can forward a comment (text, author name, link) by email to the workspace's addresses. Project Radar does this only when a team member clicks; the emails stay with the recipients and are not deleted under the limits in section 9.5.
- Government authorities, only upon a lawful request and to the extent required by law.
We do not share Google user data with anyone else: not with advertisers, ad networks, data brokers or other Project Radar workspaces.
9.5. Retention and deletion
- Refreshed at least every 30 days. The connected channel's data (title, handle, link, avatar, subscriber count), comments and replies, including the channel's own replies, video details and commenter profiles (display name, channel ID, channel link, avatar link) received through the YouTube Data API are kept while the channel is connected and the workspace uses Project Radar, and they are refreshed: 20 days after they were received, Project Radar reads them again from the API by ID (
channels.list,comments.list,videos.list; comments held for review also throughcommentThreads.listwithmoderationStatus=heldForReview; a commenter profile is refreshed together with their comment). If a comment, reply or video is missing from two successful YouTube responses in a row at least 12 hours apart, Project Radar deletes it (a video that still has comments under it is reduced to its ID and link). An API error (quota, network failure) deletes nothing: Project Radar retries after 1, 2, 4 and 6 hours. Anything that could not be refreshed is deleted no later than 30 days after it was received (Project Radar deletes it on day 29); 5 days before that, the workspace administrators receive an email and the account card shows the date. AI labels (topic, sentiment, signs of spam) are stored with the comment and are wiped when its text changes; model responses in the internal cache are kept for no more than 7 days. - Disconnect, keep history (Settings → Accounts → account settings): Project Radar immediately wipes the channel's OAuth tokens and stops receiving data from it. After 6 days, and in any case no later than 7 days after disconnection, Project Radar deletes all data it received for this channel through the YouTube Data API: comments and replies, video details, the channel's title, handle, link, avatar and subscriber count, and commenter profiles. The deletion date is shown in the account card; reconnecting the same channel before then cancels the deletion. After the deletion, the workspace keeps the channel ID and the account name: the name team members gave it in Project Radar, otherwise “Канал YouTube” (“YouTube channel”) with the channel ID (for history transferred from another service, the name and profile from that history). The channel title received from YouTube is also removed from the action log.
- Access revoked or no longer valid. If Google stops accepting the channel's authorization (access was revoked in the Google account or the token is no longer valid), Project Radar marks an authorization error and deletes the channel's data in the same way: after 6 days and no later than 7 days, unless the channel is authorized again.
- Delete the account (Settings → Accounts): Project Radar immediately deletes the channel from the workspace together with its OAuth tokens, comments and video details, as well as the profiles of the authors of those comments (if the workspace still needs an author record, the data from YouTube in it is wiped right away, see below). The same applies when a workspace is deleted.
- Commenter profiles. A commenter's profile is deleted together with their last comment received from YouTube, and the saved avatar copy together with the profile. If the workspace still needs the author record for a ban, block, note, label or for comments in transferred history (see below), only the author's channel ID remains in it, and the name, channel link and avatar received from YouTube are wiped; for an author without a YouTube channel, whom the API identifies only by name, the name is wiped too.
- History transferred from another service. Comments and commenter profiles that the workspace owner transferred to Project Radar from another service (history import) were not received through Project Radar's API client. The limits in this section do not apply to them; they are kept under the rules of section 7, and Project Radar does not add YouTube Data API data to them. Channel replies from YouTube that Project Radar attached to such comments before October 1, 2026 have been deleted; text that it refreshed from YouTube back then is refreshed and deleted under the rules of this section, and the history record itself stays, marked as having its text removed.
- Statistics. Workspace statistics are calculated from stored comments, so comments deleted under the rules of this section disappear from them as well: figures for past periods are recalculated without them.
- The copy of a comment in Project Radar's notification queue is deleted together with the comment, and no later than 2 days after the notification is sent. Notifications already delivered to team members by email or Telegram, and emails that team members forwarded to the workspace's addresses (section 9.4), stay with the recipients.
- The action log keeps the first 200 characters of each reply a team member sent. It is kept while the workspace uses Project Radar and is deleted together with the workspace.
- Automatic database backups: daily backups are kept for 7 days, weekly backups for 4 weeks, so deleted data disappears from them no later than 5 weeks after deletion. Manual backups made by the operator before maintenance can stay longer (section 7).
- A comment author or a channel owner can ask us to delete their data by writing to privacy@gomon.space. We delete it and confirm by email within 7 days (section 13).
9.6. How to revoke access
- In Project Radar: Settings → Accounts, then disconnect or delete the YouTube channel.
- In your Google account: open myaccount.google.com/permissions, choose Project Radar and remove its access. After that, Project Radar can no longer receive data from the channel or act on it.
After access is revoked, Project Radar can no longer receive or refresh the channel's data: it marks an authorization error and, unless the channel is authorized again, deletes the data it received from YouTube for this channel after 6 days and no later than 7 days (section 9.5). To delete it right away, delete the channel in Project Radar or write to privacy@gomon.space.
9.7. Limited Use
Project Radar's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
In particular, Google user data:
- is used only to provide the user-facing features described in this section: the feed, replies, channel moderation, the workspace's statistics and optional AI labeling;
- is transferred to others only in the cases listed in section 9.4, which are necessary to provide these features or to comply with the law;
- is not used for advertising, including personalized, retargeted or interest-based advertising;
- is not sold and is not shared with data brokers or information resellers;
- is not used to determine creditworthiness or for lending purposes;
- is not used to develop, improve or train generalized AI or machine learning models; Project Radar does not train any models;
- is not read by anyone on the operator's side, except with the user's consent for specific data, when necessary for security purposes such as investigating abuse, or when required by law. Team members of the workspace that connected the channel see its comments in the feed, because that is what the service is for.
Project Radar does not generate images, video or audio with AI, never uses Google APIs or Google user data to create generative AI imagery, never creates non-consensual intimate imagery (NCII) and does not allow Project Radar to be used for it. Project Radar complies with the Google APIs Terms of Service; using Project Radar for NCII is prohibited by the Terms of Service.
10. Meta data (Instagram)
An Instagram profile is connected by its owner via Facebook sign-in. Project Radar requests only the permissions needed for the feed, replies and moderation: the list of pages and the Instagram profiles linked to them, comments on the profile's posts, messages sent to the profile, and replying to, hiding and deleting comments on behalf of the profile. Direct messages can be replied to only within the window that Meta allows.
Meta data is used only for these features and is not transferred to third parties, except in the cases described in section 6. Data can be deleted using the methods in section 13.
11. TikTok data
A TikTok account is connected by its owner via TikTok sign-in. Project Radar receives the account's list of videos and the comments on them, publishes replies, and hides and deletes comments upon a team member's action. TikTok data is used only for these features.
12. Your rights
You have the right to:
- find out whether we process your data and obtain a list of it;
- demand that your data be corrected, blocked or deleted if it is incomplete, outdated or processed unlawfully;
- withdraw your consent to processing;
- appeal our actions to Roskomnadzor (the Russian data protection authority) or in court.
Write to privacy@gomon.space. We will respond within 10 business days. To help us find your data, specify the platform and a link to your profile or your username.
13. How to delete data
If you are the author of a comment or message and want Project Radar to delete your data, write to privacy@gomon.space and specify the platform and a link to your profile or your username. We will delete your comments, messages and profile data from Project Radar and confirm this by reply email within 7 days.
- Deletion from Project Radar does not delete the comment itself on the social network. To remove it there, delete it on the social network itself or contact the page owner.
- Project Radar also deletes data received from YouTube on its own: no later than 7 days after the channel is disconnected or access is revoked, and anything that could not be refreshed within 30 days (section 9.5).
- The owner of a connected page can delete it together with all its data in Project Radar: Settings → Accounts → delete account. Access can also be revoked in the settings of the platform itself: for Facebook and Instagram, in the “Apps and websites” section of the account settings; for Google, at myaccount.google.com/permissions; for TikTok, in the “Security” section of the account settings, under the item about apps and services.
- A team member can ask the workspace owner to delete their account or write to us.
14. Changes and contacts
We may update this Policy. The new version is published on this page with the date at the top. If the changes are material, we will notify workspace owners in advance.
Operator: Individual Entrepreneur Lev Aleksandrovich Evseev, INN 540440564800.
Email: privacy@gomon.space.