Privacy Policy

1. Who is responsible for your data?

Interaction Bot (the bot and the website interaction-bot.com) is operated by its administrators, who decide why and how the data below is processed. You can reach us at [email protected] or through the "Support" link of the website.

2. What information is stored?

Server settings. Discord IDs (server, channel, role, category and user) and the configuration of your server: languages, channels, mirrors, flags, blacklists, text-to-speech, FAQ and speech-to-text channels, temporary voice channels, which features are enabled. This data is deleted when the bot leaves the server.

Your account. When you log in with Discord we receive your Discord ID, username and avatar (scope "identify"; the Activity needs nothing more). We store your Discord ID when you activate a premium or are blacklisted, the link between your Discord ID and your Patreon account when you use premium, the date of your last vote on top.gg, and your personal preferences (default language, public replies, vote reminder, text-to-speech voice, temporary voice channel preferences and trusted or blocked members).

Messages. If an automatic translation is set up or you ask for a translation or a language detection, the text of the message is processed in order to be translated. It is not stored in the database and does not appear in our logs, and it is not saved in any persistent storage. The translation service only keeps recently translated texts in its working memory (RAM) for a short time, as a cache to answer faster; this memory is never written to a disk and is cleared on each restart and every day.

Voice. Voice transcription (/transcribe) processes the audio in memory, transcribes it, then discards it; the resulting text is posted in the text channel chosen by the server and shown in the live view (the Activity). An audio recording can keep voice data, and only when a member of the server starts it:

• Audio recording (/record): one audio track per speaker plus a mix of the voice channel, kept encrypted for 24 hours (7 days on premium servers), then deleted. Everyone in the channel is warned by a message in the text channel, a spoken announcement and the nickname of the bot ("Recording") for as long as it records. Only the member who started a recording can download it. The notes added during a recording (text, author, time) are stored with it. The voice of the bot itself (text-to-speech) can be recorded on its own track when the server enables it.
• Transcription of a recording (on request): only when the member who started a recording asks for it, each track is sent to our speech-to-text service in a background job and the resulting text (with the name of each speaker) is stored encrypted with the recording. Only that member can read it; it is deleted with the recording and by "Delete my data".

Voice cloning (premium, optional): if you create a cloned voice, the voice sample you provide is used to create it, with your explicit consent shown before the creation. A cloned voice is deleted when you delete it, and automatically 30 days after the premium it requires ends.

AI configuration assistant (optional, for administrators): when you use the assistant of the dashboard, or the /ai setup command (also started from the button of /help), the messages you write to it, the earlier messages of the same conversation and the server settings it reads or changes (names of channels, roles and categories, languages, FAQ entries, etc.) are sent to a third-party AI model provider to produce the answer (see section 4). With /ai setup the bot reads only the messages of the administrator who started the conversation, in the channel where it was started, until the conversation ends: the Stop button, five minutes without a message, the daily limit of the server, the assistant judging the work done, or the loss of the Manage Server permission. Messages of other members are never read, and a member who opted out of message processing (section 7) cannot use it. We do not keep the conversation once it ends. We keep a log of the changes made (the operation, a short description, the result, the server and user IDs, the date) and a daily request counter per server.

Invitation tracking (optional, off by default): when an administrator turns it on, the bot records, for each new member, the invitation used and who created it, the dates of arrival and departure, and counters per inviter, for the period the server chose (30 to 365 days). A member who opted out of message processing is counted for the inviter but is not named in the journal.

Honeypot (optional, off by default): when an administrator turns it on and chooses a trap channel, a member who writes in it is punished as the server decided (kick, ban or timeout). We then keep, for 30 days, the ID of the member, the date, the action and its result, the date their account was created and the date they joined, and the first 300 characters of the message as evidence, visible to the administrators of the server.

Welcome, leave and private messages (optional): the messages and banners use the name, the avatar and the position in the server of the member concerned. The pictures an administrator uploads for a banner are stored on our own storage (up to 10 per server) until they delete them or the bot leaves the server.

Opt-out. If you opt out (see section 7), we keep your Discord ID in a list so that the bot never records your voice again.

Anonymous technical statistics (performance and load metrics), collected on our own servers, never contain user, server or channel IDs, nor the content of messages.

3. Why we store this information, and on which legal basis

We use this data to provide the services you ask for (performance of the service), to secure and operate them (our legitimate interest: abuse prevention, limits, reliability) and, for recordings and cloned voices, on the basis of the information given to everyone in the channel and of your consent where it is asked (for example the explicit consent before a voice is cloned). You can object at any time with the opt-out. We never sell your data and never use it for advertising or to train models.

Connections to our website use HTTPS. The website only uses the cookie that keeps you logged in (strictly necessary); it has no advertising or analytics cookie.

4. Who gets this data?

The data is only accessible to the administrators of Interaction Bot. Translation, language detection, text-to-speech, speech-to-text and voice cloning are performed by our own models, on our own servers: no third-party AI provider receives your messages or your voice. The only exception is the optional configuration assistant described in section 2: the messages written to it, and the settings it reads, are sent to an AI model provider (OpenRouter, which forwards the request to the host of the model) to be answered, under that provider's own privacy policy. Nothing else you write on Discord is sent to it.

We use the following third-party services: Discord (the platform the bot runs on), Patreon (premium subscriptions), top.gg (votes), OpenRouter and the model hosts it uses (the configuration assistant only), Cloudflare (network, DNS and protection of the website) and the hosting providers of our servers (OVH and PulseHeberg). Some of these providers can process data outside the European Union; where they do, they rely on the safeguards required by the GDPR (standard contractual clauses or an adequacy decision). No other data is shared with third parties.

5. How long we keep it

• Audio recordings: 24 hours (7 days on premium servers), then deleted automatically.
• Transcriptions of recordings: with the recording they belong to (24 hours, 7 days on premium servers).
• Text transcripts of voice channels saved by an old version of the bot (this feature no longer exists): deleted automatically within 7 days.
• Cloned voices: until you delete them, or 30 days after the end of the premium they require.
• Server settings: until the bot leaves the server. Account preferences, votes and the Patreon link: until you delete your account or data.
• Honeypot catches (with the evidence): 30 days.
• Invitation journal: the period chosen by the server (30 to 365 days); the counters per inviter until the server resets them or the bot leaves.
• Messages sent to the AI configuration assistant: not kept by us after the conversation; the log of the changes it made and the request counters: 90 days.
• Opt-out list: until you cancel your opt-out.
• Encrypted backups of our database: kept for a limited time, then overwritten; data you deleted disappears from them when they are renewed.

6. How we protect it

Recordings are encrypted, and transferred over HTTPS. An audio recording can only be downloaded by the member who started it. Access to our servers is restricted to the administrators. If a breach affecting your data ever occurred, we would inform you and the competent authority as the law requires.

7. Your rights, and how to use them

Under the GDPR you can access your data, receive a copy of it, have it deleted, object to its processing, restrict it and withdraw your consent. Log in on the website and open Account → Privacy:

• Download my data: a ZIP with your account data and preferences, your own voice tracks, and the notes you wrote (the voices and words of other people are not included, to protect them).
• Delete my data: removes your voice from the recordings (the mix is rebuilt without it), the recordings you started, your notes, your preferences and your cloned voices.
• Delete my account: everything above plus the link with your Patreon account (your Patreon subscription itself is managed on Patreon).
• Stop recording me (opt-out): the bot no longer records your voice; you can cancel it at any time. The command /privacy explains it in Discord.
• Do not process my messages (opt-out): the bot no longer translates, reads aloud or analyses your new messages, and refuses to translate them by saying that you opted out; you can cancel it at any time. The AI configuration assistant (/ai setup) then refuses to read your messages. Messages already processed are not changed.

You can also write to [email protected]; we answer within one month. If you think your rights are not respected, you can complain to your data protection authority (in France, the CNIL).

8. Children

Discord requires its users to be at least 13 years old (or the minimum age of their country). The bot is not aimed at younger children and we do not knowingly keep their data; contact us if you believe a child's data is stored, and we will delete it.

9. Questions and changes

If you have any question about the stored data, write to us (section 1) or to an administrator (role "Administrator") on the support server, which can be known via the command "/support". You can remove the bot from your server at any time; the configuration data of that server (settings, channels, roles, mirrors, flags, blacklists and premium server status) is then deleted automatically.

We may update this policy when the bot changes; the date above shows the current version and the date it takes effect, and we announce significant changes on the support server.