Privacy Policy
Last updated: July 27, 2026
This document is provided in English.
This Privacy Policy explains how Responses collects, uses, stores, shares and protects personal data when businesses use Responses to manage messages from official social, messaging and email channels, including Instagram, Facebook Messenger, WhatsApp Business, Telegram Business, Booking.com, Airbnb, Gmail and Outlook. It also explains the choices available to workspace owners, team members and the customers whose messages are processed through the service.
References to "Responses", "we", "us" and "our" mean the operator of the Responses service. References to "Customer" mean the business or organization that creates a Responses workspace. References to "End User" mean a person who sends messages to a Customer through a connected channel.
1. Our Role
For workspace account data, billing data and product usage data, Responses acts as a data controller or business. For message content, connected channel credentials and End User data processed on behalf of a Customer, Responses acts as a service provider or processor. The Customer controls which channels are connected, which team members have access, whether AI is enabled, whether replies are sent manually or automatically, and how long business records are retained within the service. Our Data Processing Addendum applies when Responses processes personal data on a Customer's behalf.
2. Data We Collect
We collect the following categories of data:
- Account data: names, email addresses, password hashes, workspace names, roles, team membership, preferences and authentication/session records.
- Connected channel data: channel type, non-secret account identifiers, display names, connection status, webhook timestamps and encrypted access credentials needed to operate official platform APIs.
- Message data: inbound and outbound messages, conversation metadata, message timestamps, delivery status, customer display names and platform-provided customer identifiers. We process attachments as platform references; we do not download or archive media files unless a product feature expressly requires it.
- AI assistance data: prompt context assembled from conversation history, workspace instructions, knowledge base entries, draft replies, confidence signals, handoff decisions and audit information about approval or auto-send actions.
- Billing data: plan, subscription status, billing email, payment processor customer identifiers, invoice references and payment status. Full card numbers are handled by our payment processor and are not stored by Responses.
- Technical and security data: IP-derived security signals, device/browser information, request metadata, rate-limit events, login/security events, service logs, cookie/session identifiers and diagnostic records.
- Support and communications data: messages sent to us for support, feedback, abuse reports, deletion requests or legal notices.
3. Sources of Data
We receive data from:
- Customers and team members who create accounts, configure workspaces or contact us.
- Official platform APIs and webhooks after a Customer connects a channel, including Meta, WhatsApp Cloud API, Telegram Bot API/Telegram Business updates, the Google Gmail API (Gmail), Microsoft Graph (Outlook), and Channex acting as the connectivity layer for Booking.com and Airbnb messages.
- Payment processors and infrastructure providers that support billing and service delivery.
- Security, hosting and analytics systems that generate operational metadata.
4. How We Use Data
We use data to:
- Provide the inbox, team workspace, channel connection and messaging features.
- Receive, normalize, store and display customer conversations.
- Generate AI draft replies, classification, handoff signals and conversation summaries.
- Send approved, automated or agent-authored replies through official APIs.
- Authenticate users, enforce roles and protect workspaces from unauthorized access.
- Operate billing, subscriptions, usage limits, support and service communications.
- Detect abuse, debug service issues, enforce our Terms and comply with legal obligations.
- Improve service reliability, security and product quality using aggregated or de-identified data.
5. AI Processing
Responses uses AI to draft replies and assist support workflows. AI output may be inaccurate, incomplete or inappropriate, so Customers remain responsible for reviewing configuration, approving replies where required and monitoring automated behavior. By default, AI starts in draft mode unless the Customer enables auto-send.
- We do not use Customer message content or End User data for advertising.
- Responses supports OpenAI, Google Gemini and Anthropic Claude as AI subprocessors. The service sends each request to only the AI provider active for its deployment. We do not use Customer message content or End User data to train our own general AI models, and we do not opt in to provider model training or improvement.
- We send the active provider the recent conversation needed for the reply, the active assistant instructions and selected knowledge base snippets. Before transmission, Responses masks detected email addresses, phone numbers, payment-card numbers and IBANs in each of those text sources. Names, postal addresses, free-form identifiers and other personal data may remain in the selected context.
- Every OpenAI Responses API request sets
store: false, which disables Responses API application-state storage for that request. This is not Zero Data Retention: OpenAI's default abuse-monitoring logs may contain inputs and outputs for up to 30 days, subject to its documented legal and safety exceptions. - Every Gemini Interactions API request sets
store: false, which disables Interaction-object storage. Production uses a paid project without log sharing. Google states that paid-service prompts and responses are not used to improve its products; limited abuse-monitoring processing may still apply unless Zero Data Retention is approved for the project. - Claude requests use Anthropic's stateless Messages API without Files, prompt caching, tools or feedback. Anthropic states that commercial API inputs and outputs are not used for model training by default and are deleted from its backend within 30 days under standard retention, subject to Usage Policy, legal and contractual exceptions.
- The application is designed to avoid logging message content, prompts, access tokens or generated replies in operational logs.
6. Telegram Business, Channex and Connected Platforms
When a Customer connects Telegram Business, the Customer enables Secretary Mode for the Responses bot in Telegram. Telegram then sends business connection and business message updates to Responses. Responses uses the business connection identifier to receive customer messages and send replies on behalf of the connected Telegram Business account, according to the permissions granted in Telegram.
When a Customer connects Booking.com or Airbnb through Channex, the Customer supplies a Channex API key and property identifier. Responses validates the property, stores the API key encrypted, and registers an authenticated Channex webhook. Channex relays guest messages, message-thread metadata and attachment references from the selected OTA and carries replies back to that thread. Airbnb inquiries can create a thread without a booking. Disconnecting deletes the stored API key and asks Channex to remove the webhook; conversation records already stored in Responses remain subject to the Customer's Responses retention settings.
Customers are responsible for their relationship with each connected platform, including platform terms, messaging policies, consent/opt-in requirements, customer notices and any restrictions on automated replies.
7. Email Channels: Google Gmail and Microsoft Outlook
When a Customer connects a Gmail mailbox, they authorize Responses through Google OAuth to read incoming messages and send replies on their behalf. Responses requests only the narrowest scopes needed to operate a shared inbox: gmail.readonly to read the incoming customer emails a reply is generated for, and gmail.send to send the reply into the same email thread. Responses registers a Gmail watch so Google notifies the service of new inbox messages; the notification itself contains no message content, and Responses then fetches the referenced messages using the authorized scope. When a Customer connects an Outlook mailbox, the equivalent Microsoft Graph delegated permissions (Mail.Read, Mail.Send, User.Read, offline_access) are used the same way.
Responses uses Gmail and Outlook data solely to provide the user-facing inbox and AI-assisted reply features described in this policy. It does not use this data for advertising, does not sell it, does not transfer it to others except the subprocessors required to run the features (see Section 10), and does not allow humans to read it except where you or your team explicitly view a conversation, where required for security or to comply with law, or on data that has been aggregated or de-identified. Disconnecting a mailbox revokes the tokens, stops the Gmail watch or Graph subscription, and deletes the stored credentials.
8. Google API Services User Data Policy (Limited Use)
Responses' use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including its Limited Use requirements. Specifically, Google user data obtained through Gmail scopes is used only to provide and improve the user-facing features that are prominent in the Responses interface (the shared email inbox and AI-assisted replies); it is not used for serving advertising; it is not sold; it is not transferred to others except as necessary to provide or improve those features, to comply with applicable law, or as part of a merger, acquisition or asset sale with user notice; and humans do not read it except with the user's or workspace's explicit consent for viewing their own conversations, for security purposes such as investigating abuse, to comply with applicable law, or on aggregated/anonymized data. Responses does not use Google user data to train generalized AI/ML models.
9. Cookies and Similar Technologies
Responses uses cookies and similar local storage technologies for authentication, session continuity, security, preferences and product functionality. We do not use message content for behavioral advertising. Browser controls may allow users to delete or block cookies, but doing so may prevent login or workspace functionality from working correctly.
10. Sharing and Subprocessors
We share data only as needed to provide, secure and support the service:
- OpenAI, Google and Anthropic may process the minimized and redacted context described in Section 5 through the single AI provider active for the deployment, to generate draft replies and related AI outputs.
- Channex processes Booking.com and Airbnb message content, guest and property identifiers, message-thread metadata and attachment references to provide the OTA connectivity selected by the Customer. Channex's current terms and privacy policy and documented retention periods apply to its service.
- Payment processors process billing events, invoices and payment information.
- Cloud, database, queue, logging and security providers host and protect the service.
- Connected messaging platforms receive outgoing replies and provide inbound webhook events as configured by the Customer.
- Professional advisers and authorities may receive data where necessary for legal compliance, safety, fraud prevention, tax, accounting or dispute resolution.
We do not sell personal data or message content.
11. Retention
We retain data for as long as needed to provide the service, comply with legal obligations, resolve disputes, enforce agreements and maintain security. Workspace conversation retention may be controlled by product settings. Connected channel tokens are deleted when an account is disconnected. Deletion requests remove workspace conversations, messages, connected accounts, platform credentials and related business records except where limited retention is required by law, security, backups or billing records. AI provider handling of request data is described separately in Section 5.
12. Security
Responses uses technical and organizational measures designed to protect data, including encrypted transport, role-based workspace access, encrypted storage of platform credentials, secret-token webhook verification, audit/security events and restricted operational logging. No system is perfectly secure. Customers must protect their credentials, restrict team access, rotate exposed platform tokens and notify us promptly about suspected compromise.
13. International Processing
Responses, its infrastructure providers and subprocessors may process data in countries other than where the Customer or End User is located. Where required, Customers and Responses will rely on appropriate transfer mechanisms, contractual commitments and subprocessors that support lawful international processing.
14. Privacy Rights
Depending on location, individuals may have rights to access, correct, delete, export, restrict or object to processing of personal data, and to appeal or lodge a complaint with a privacy authority. Team members can use in-product controls where available. End Users should first contact the business they messaged, because that Customer controls the conversation. If an End User contacts Responses directly, we may route the request to the relevant Customer unless law requires us to respond directly.
15. Meta User Data Deletion
Responses supports Meta's user data deletion callback. If a person removes the Responses app or requests deletion through Meta's app settings, Meta can call our deletion endpoint at https://responses.me/api/v1/meta/data-deletion. The callback verifies Meta's signed request, deletes Facebook/Instagram data associated with the app-scoped Meta user ID, and returns a confirmation code with a status URL.
Step-by-step deletion instructions and a status checker are available at responses.me/data-deletion. People can also request deletion by contacting privacy@responses.me. Users can delete a workspace or their entire account (including cancellation of active subscriptions) from the in-product Security & Privacy page.
16. Children
Responses is a business service and is not intended for children. Customers must not use Responses to knowingly collect personal data from children unless they have all required permissions and legal authority to do so.
17. Changes
We may update this Privacy Policy from time to time. If changes are material, we will provide notice through the service, email or another reasonable method. Continued use after the effective date means the updated policy applies.
18. Contact
Privacy requests, security notices and data deletion requests can be sent to privacy@responses.me. General questions and support: support@responses.me. Workspace owners can also use the in-product Security & Privacy controls to export or delete workspace data.