Privacy Policy

Last updated: 4 October 2026

1. Who we are

The dmhive service (dmhive.io) is provided by Individual Entrepreneur Valentin Nazarenko, registered at Georgia, Tbilisi, Nadzaladevi District, Tsotne Dadiani Street, N34, Building N7, Block N10, Commercial Space, registration number 300421933 ("we", "us", "the service").

Data protection enquiries: privacy@dmhive.io. Phone: +995 591 06 95 19.

2. Two roles: our customers and their subscribers

• Customer: a business that signed up for dmhive and connected its messaging accounts. For customer data we act as the data controller.
• Subscriber: a person who messages the customer. For subscriber data we act as a data processor, processing it on the customer's documented instructions under a data processing agreement. The customer is the controller of that data.

3. Data we process

3.1. Customer account data

• email address, name, password (stored only as an irreversible hash);
• subscription and payment history (card details are handled by our payment provider Paddle, see section 6, and never reach our servers);
• technical data: IP address, browser type, activity logs.

3.2. Data received from Meta platforms

When a customer connects an Instagram account or Facebook Page, we receive the following through official Meta APIs:

Data	Purpose
Connected account ID, name and profile picture	Show the customer which account is connected; send messages on its behalf
Direct messages: text, attachments, sender ID, timestamps	Deliver automated replies, display conversations in the inbox, let operators reply
Comments on posts and story replies	Trigger keyword-based automations, including replying to a comment via direct message
Subscriber public profile: name and picture	Identify the person in the inbox and personalise replies
Access tokens	Technical ability to send and receive messages; stored encrypted

We receive equivalent data from Telegram and WhatsApp, limited to what the bot needs to operate in that channel.

3.3. Chat archives uploaded by the customer

A customer may upload an export of their own chats so that the service can assemble a bot scenario automatically. Such archives receive our strictest handling:

• media files are discarded during extraction and never enter processing;
• known fields are stripped from the text before any analysis: phone numbers, emails, links, long numbers and participant names are replaced with anonymised placeholders; free-form addresses and names within the message text may remain;
• a source archive uploaded inside a customer account is kept for no longer than 7 days from upload and is then deleted automatically; during that time it is used to build the scenario and to check and improve the quality of analysis (section 4); it is deleted immediately when the account is deleted; for the no-signup demo the archive is deleted right after processing;
• the language model used to build the scenario is OpenAI (section 6), which receives only an anonymised distilled excerpt, never the full archive.

3.4. Payment receipts sent by buyers

If a customer's bot has asked for payment for an order in the chat, the buyer may reply with a screenshot or PDF of the receipt. We read that file to check the order payment, and we keep it so that the customer's account owner can see which exact file the payment was accepted on or passed to them for review. The file may contain the payer's and recipient's names, part of an account number, the amount and the date of payment. We act on the customer's instructions, as with the rest of the conversation.

• only images and PDFs sent in reply to a payment request in the customer's bot chat are stored, and only after acceptance of this policy as revised on 4 October 2026; we do not store other attachments from buyers: the dashboard downloads them from the platform when you view them;
• the file is encrypted (AES-256-GCM) and kept on the disk of a server in Germany; it is not included in the daily database backups;
• only the customer's account owner can see the file, on the order card in the dashboard; staff with the operator role cannot;
• retention: while the order is open and for 30 days after it is closed (paid or cancelled), but no longer than 90 days from receipt; a receipt that arrives without an open order is kept for 30 days (section 8);
• the file is deleted together with the contact and with the customer's account, and together with an Instagram contact when Meta sends a data deletion request for that person; in all other cases it is deleted when its retention period ends.

3.5. Conversations the customer marks for review

A customer may mark inbox conversations where the bot fell short and ask the service to suggest additions to the scenario. In that case:

• only the marked conversations from the 14 days before marking and the customer's note are sent to the OpenAI language model (section 6), after phone numbers, email addresses, links, long numbers and the contact's name as the channel spells it are removed;
• the model suggests price list entries, reference entries and menu buttons; they enter the scenario only if the customer accepts them, and take effect in the channel after publishing;
• marks and suggestions are kept no longer than inbox conversations (section 8), and no models are trained on these conversations.

4. Why we process data

• providing the service: receiving and sending messages, running scenarios, inbox, contacts;
• checking payment for an order against a receipt the buyer sent in the customer's bot chat;
• automatically assembling a bot scenario from material the customer provided;
• extending the bot scenario from conversations the customer marked for review;
• checking and improving the quality of analysis on uploaded archives within their retention period (7 days): we measure how fully the service finds questions, prices and facts, and adjust our own analysis instructions; once the archive is deleted, only anonymised findings remain: statistics and typical questions without names, contact details or other data that could identify a person;
• billing and plan usage accounting;
• customer support;
• security: preventing abuse, spam and unauthorised access;
• compliance with legal obligations.

GDPR legal bases: performance of a contract (Art. 6(1)(b)); legitimate interests in operating and securing the service (Art. 6(1)(f)); consent where required (Art. 6(1)(a)); legal obligation (Art. 6(1)(c)).

5. What we do not do

• we do not sell data or share it for advertising purposes;
• we do not train artificial intelligence models on conversation content and do not share it for training; the provider that assembles the scenario (OpenAI) does not use data for training under its API terms; the quality checks in section 4 measure and adjust our analysis instructions and are not model training; input to the AI-node simulator of the public demo is processed by Google Gemini on the free plan, and Google may use that input to improve its own services; simulator conversations in customer projects are not sent to Google;
• we do not read customer conversations manually, except when the customer asks for help, when required by a security investigation or by law, and for spot checks of quality-check results within the archive retention period (section 4).

6. Who we share data with (subprocessors)

• infrastructure and hosting: Hetzner Online GmbH, Germany;
• content delivery, site protection and storage of early-access requests: Cloudflare, Inc.; early-access form submissions are kept for up to 12 months; Cloudflare R2 (EU jurisdiction) also stores our daily database backups, encrypted with our key before upload, so Cloudflare cannot read them;
• language model provider for building the scenario, processing the owner's notes and analysing the conversations the owner marks, after anonymisation: OpenAI, L.L.C., USA; under its API terms data is not used for training, abuse logs are retained for up to 30 days;
• model provider for AI replies in the public demo simulator: Google LLC (Gemini API), USA, free plan;
• app email delivery (address confirmation, password recovery, notifications): Zoho Corporation, Zoho Mail, servers in the EU;
• delivery of bot messages and owner notifications in Telegram: an independent platform with its own terms;

Payments are handled by Paddle.com Market Limited (United Kingdom), acting as our reseller (merchant of record) and an independent controller of payment data. Paddle receives the buyer's name, email address, country and payment details; full card details never reach our servers. See the Paddle privacy policy for details.

An up-to-date subprocessor list is set out in our Data Processing Agreement.

Requests from public authorities. We disclose personal data to public authorities only when the law requires it. Before responding, we check every such request for legality and validity, disclose only the minimum data needed to comply, and keep a record of the request, our response and its legal basis.

7. Data location and international transfers

Primary storage is located in the European Union, in Germany, and encrypted backups are kept in the EU with Cloudflare R2. A copy of an uploaded archive used for quality checks is processed on a dmhive work computer in Georgia and deleted within the same period as on the server, no later than 7 days from upload. Some subprocessors process data outside the EU: an anonymised summary of conversations and the conversations the owner marks, after anonymisation, are sent to OpenAI (USA) for analysis, and public demo simulator input is sent to Google (USA). Early-access form submissions are stored with Cloudflare (USA, EU). Such transfers rely on the European Commission's Standard Contractual Clauses or other safeguards permitted by law.

8. Retention

Category	Retention
Uploaded chat archives	up to 7 days from upload, on the server and on the work computer; immediately when the account is deleted; for the no-signup demo, right after processing
Inbox conversations	90 days from the message, then deleted automatically; immediately when the account is deleted
Screenshots and PDFs of payment receipts that buyers send to customers' bots	while the order is open and for 30 days after it is closed (paid or cancelled), but no longer than 90 days from receipt; a receipt without an open order, 30 days; encrypted, not included in backups; immediately when the contact or the account is deleted
Conversation marks for review and scenario suggestions	up to 90 days, together with inbox conversations; immediately when the account is deleted
Contact records	while the account is active; immediately when it is deleted
Customer account data	while the account is active; anonymised immediately on deletion, except the records below
Payment and accounting records	as required by law
Technical logs	up to 90 days
Backups	up to 35 days, then overwritten
Records of document acceptance and data requests	kept as proof that we met our obligations
Images added to the bot scenario	public via link for as long as the account exists
Early-access form submissions	up to 12 months

9. Your rights

You may request access to your data, its correction, deletion or portability, restrict or object to processing, and withdraw consent previously given. You can download a copy of your data and delete your account yourself in the "Data and requests" section of the app. By email to privacy@dmhive.io we respond no later than 10 working days from receiving the request, as required by the Georgian Law on Personal Data Protection; if the law allows an extension for a complex request, we will tell you about it and the reasons before that deadline expires.

If you are a subscriber who messaged a business using dmhive, please contact that business first: it decides how your data is used. We will assist it in fulfilling your request. See Data Deletion for the deletion procedure.

You also have the right to lodge a complaint with your local data protection authority.

10. Security

Data is transmitted over encrypted connections; access tokens are stored encrypted, passwords only as an irreversible hash; staff access follows the principle of least privilege. Customer workspaces are isolated from one another. In the event of a personal data breach we notify affected customers and the supervisory authority within the deadlines set by applicable law (72 hours under the GDPR).

11. Cookies

dmhive.io uses strictly necessary cookies only: login session and interface preferences. There are no advertising or third-party tracking cookies on this site.

12. Children

The service is intended for businesses and is not directed at persons under 16. We do not knowingly collect their data.

13. Changes to this policy

We may update this policy. We notify customers of material changes by email and in the product interface at least 14 days before they take effect. The date of the current version is shown at the top of this page.