Darto Privacy Policy

Last update: 2026-11-18
Owner: David Chicano
Contact email: help@darto.es
Country: Spain

Darto is a SaaS platform for businesses that automates customer communication and operational workflows through messaging channels and optional integrations. This policy explains what data we access, how we use it, who we share it with, how we protect it, and how users can request deletion.

1. Data we process

1.1 Account and business data

  • Business account name, email address, and login data.
  • Business information entered by the user or publicly available, such as business name, address, opening hours, or descriptions.
  • Assistant settings, connected integrations, and operational preferences.
  • Documents the business uploads as knowledge for its assistant, and the text, sections, and summaries Darto extracts from them. They are meant for business information: we recommend not uploading documents that contain customers’ personal data.

1.2 Messaging and service usage data

  • Messages sent and received through connected channels such as WhatsApp Business and Instagram.
  • Operational and technical data needed to run the service, including logs, IP addresses, internal metrics, startup timings, and error records. Mobile metrics may include a pseudonymous installation identifier, device details, app version, and installed update. We do not add conversation content or customer or business identifiers to performance metrics; automatically captured errors may include a technical description of the failure.

1.3 Optional integration data

If the business owner chooses to connect a third-party integration in Darto, we may access and process the minimum data needed to configure that integration and provide the requested functionality. Depending on the integration, this may include account or business identifiers, configuration data, availability data, operational data, and other information needed to query or perform actions authorized by the user.

1.4 Website visits

To learn how our website (darto.app and darto.es) is used, we record what happens during each visit: the pages viewed, how far the visitor scrolls, the buttons and links they press, and whether the opening animation is watched to the end or skipped. Each visit carries a random identifier that the page creates and keeps only in memory: it disappears when the tab is closed or reloaded and cannot recognise the visitor on another visit. This measurement uses no cookies and stores nothing in the browser or on the device, and we do not record the IP address. Alongside the events we keep only general data: approximate country, device type, window size, language, the referring website and the link’s campaign tags. We do not combine this data with user accounts or the early access list. The data is stored in Supabase and deleted after 13 months.

2. How we use data

We use the above data only for the following purposes:

  • Create and manage the business account in Darto.
  • Provide the contracted product functionality.
  • Process customer conversations and generate AI-assisted replies or actions.
  • Connect, maintain, and reconfigure third-party integrations authorized by the business.
  • Query or perform operations in the relevant integration only when the user has chosen to connect it and the requested functionality requires it.
  • Display synchronized information to the business within the app.
  • Detect failures, prevent abuse, maintain security, and troubleshoot technical issues.
  • Comply with legal and regulatory obligations.

Third-party integrations are used only when the business chooses to connect them and authorizes their use within Darto.

3. Data sharing and third parties

We share data only when necessary to provide the service or when required by law.

3.1 Service providers and processors

  • Supabase: infrastructure, database, and operational storage, including the website visits described in section 1.4.
  • Cloudflare: website hosting and delivery, and Cloudflare’s own visitor analytics, which use no cookies.
  • Google Sign-In: account sign-in and identification when you choose Google authentication; its SDK may process identifiers and technical data for functionality and analytics.
  • Sentry: error diagnostics, performance monitoring, and session replay to troubleshoot technical issues. Session replay is configured to mask text, images, and vectors. We enforce organization-level data scrubbing, prevent storage of IP addresses in new events, and limit notification detail. These controls do not guarantee that free text is anonymous or delete earlier events.
  • Expo (EAS Observe): technical mobile-app performance and startup metrics associated with app releases and updates. The version in use may also record unhandled errors and native crashes; Sentry remains our primary system for investigating them. We do not send business events or route parameters identifying customers, conversations, or businesses to Observe.
  • OpenAI and Google Gemini: conversation processing and assistant features through their APIs, including reading and summarizing the documents the business uploads, and failure recovery. OpenAI also transcribes the audio when the user dictates a message.
  • Anthropic: replies of Darto’s internal agent, which reads account data (settings, flows, requests, and client conversations) to answer the user; OpenAI acts as fallback. The assistant’s replies to clients do not go through Anthropic.
  • xAI: available for specific diagnostics configured by Darto; it does not receive ordinary conversations.
  • TypeSafe AI (Jev): bounded decisions through its direct API. It evaluates the actions the assistant proposes to run without manual approval, using relevant conversation excerpts, business or integration context and the action parameters. It also classifies the messages the user writes to the internal agent (the message and up to four of their previous ones, without Darto’s replies) to choose the model that answers. Darto does not add integration credentials to that input. Vercel AI Gateway is not an intermediary on this route.
  • Meta Platforms: connected messaging channels such as WhatsApp Business and Instagram.
  • Firecrawl: reading the public pages of the business’s website when Darto analyzes it (filling in the editor from the website, during sign-up or when connecting a Shopify store), and converting the documents the business uploads for its assistant into text. Darto does not have Firecrawl’s zero data retention: Firecrawl may temporarily keep logs and cached copies under its own terms.
  • The third-party provider or external service corresponding to the integration connected by the business, only when needed to query data or perform the requested functionality.

We select shared context according to each provider’s purpose. Conversation excerpts may contain personal data; limiting context does not automatically anonymize free text.

3.2 What we do not do

  • We do not sell personal data or data obtained through third-party integrations to third parties.
  • We do not share integration data with third parties for advertising purposes.
  • We do not access a third-party integration without prior authorization from the relevant user or business.

We may also disclose information if necessary to comply with a legal obligation, respond to a valid request from a competent authority, or protect the security of the service.

4. Data storage and protection

We apply reasonable technical and organizational measures to protect personal data and data processed through integrations:

  • Encrypted transmission over HTTPS/TLS.
  • Storage on Supabase-managed infrastructure with provider security controls.
  • Access controls at the application and database level.
  • Restricted access to systems and credentials only for authorized service components.
  • Data minimization and processing limited to the purposes described in this policy.
  • Server-side storage of integration credentials or tokens when needed to maintain an authorized connection; they are not exposed to the business’s end customers.
  • Temporary mobile-app to web-app session handoff through opaque one-time tokens, without placing Supabase session tokens in URLs.

5. Retention, disconnect, and deletion

We retain data for as long as needed to provide the service, keep an integration active, troubleshoot issues, comply with legal obligations, or handle legitimate claims. Website visits (section 1.4) are deleted automatically after 13 months.

For third-party integrations generally:

  • While the integration remains active, Darto retains the connection and configuration data needed to operate that integration.
  • If the business disconnects an integration from within the Darto app, Darto deletes the integration link and associated synchronized data needed for that connection, and Darto stops accessing that integration from that point onward.
  • Users may also request deletion of their data by emailing help@darto.es.

We may retain limited technical logs or backup copies for a reasonable period when required for security, fraud prevention, service continuity, or legal compliance.

Under the current Sentry plan, new errors, logs, traces and replays are retained for 30 days. Data sent during earlier plans or trials may keep different retention periods. Expo applies a 90-day retention window to EAS Observe metrics and errors on all plans. Deletion runs monthly, so some records may remain after those 90 days. Early deletion is handled through Expo support; deleting the Expo project does not itself delete its Observe data. According to Expo, database backups are kept for a short period and expire on their own schedule, with no exact period confirmed. Turning off ingestion prevents Observe from accepting new data but does not delete data already received. Deleting a Darto account does not immediately delete every technical copy held by these providers. You can contact us to coordinate applicable deletion requests.

We process personal data on the following legal bases:

  • Performance of the contract with the business user.
  • User consent when connecting services or third-party integrations that require authorization.
  • Legitimate interest for security, abuse prevention, technical maintenance, and service operations, including learning how the website is used (section 1.4).
  • Compliance with legal obligations.

7. User rights

You may request access, rectification, deletion, restriction, objection, or portability of your data by writing to help@darto.es.

You may also revoke access to an integration by disconnecting it from Darto and, where the provider supports it, later revoking the app’s access from your account with that provider.

8. International transfers

Some providers used by Darto may process data outside the European Economic Area. In those cases, we apply reasonable safeguards, including standard contractual clauses or other appropriate protections where applicable.

TypeSafe states US hosting. Its Data Processing Addendum incorporates EU standard contractual clauses, including processor-to-subprocessor transfers where applicable. Linking the provider’s public terms does not itself establish a specific agreement for Darto.

9. AI processing

When an assistant feature needs to use business data, conversation data, or data from a third-party integration to answer a request or perform an authorized action, Darto selects and processes context to interpret the request, retrieve relevant information, and return a useful response.

Darto does not sell data obtained through third-party integrations and does not use it for advertising.

Before using Darto, the app asks each user to accept a notice stating what data is sent to these providers and why; Darto records the date and version of the accepted notice. If the providers or the kind of data sent change, the notice is shown again.

Voice dictation: the audio is sent to OpenAI for transcription and Darto does not store it; it only keeps the resulting text when the user sends it.

Darto does not use this data to train generalized models. OpenAI, Anthropic, Google, and TypeSafe do not train their models on data sent through their paid APIs under their commercial terms. TypeSafe’s Master Customer Agreement excludes model-weight training on customer data without prior consent. This does not mean zero retention: service, security, billing, legal and telemetry processing may remain. Public terms do not specify a fixed API retention period; we do not claim enterprise ZDR for Darto. MiMo/Xiaomi and DeepInfra remain outside ordinary customer-facing routes; testing uses synthetic data.

10. Provider-specific information

Some integrations or providers may require additional, provider-specific information about the data accessed, how it is used, how it is stored, or how it is deleted. In those cases, Darto may publish provider-specific annexes or sections without changing the general structure of this policy.

10.1 Google Calendar

If the business owner chooses to connect Google Calendar through OAuth, Darto may access and process the following categories of Google user data:

  • The email address of the connected Google account.
  • The list of calendars available in that account during setup, including identifiers, names, descriptions, time zone, access role, and whether a calendar is primary.
  • The calendar selected by the user for use with Darto, including its ID and name.
  • The OAuth permissions granted and the tokens required to keep the connection active.
  • Availability data for the selected calendar through free/busy queries.
  • Event data from the selected calendar when needed to provide the requested functionality, including title, description, location, start and end date/time, time zone, event status, recurrence, and, when present on the event itself, attendee or organizer information.

Darto uses this Google data only to:

  • show available calendars during setup;
  • let the user select the calendar that will be connected to Darto;
  • check availability;
  • list or retrieve events;
  • create, update, or cancel user-authorized events;
  • respond inside Darto to appointment-management requests.

Darto does not request access to Google services unrelated to the integration authorized by the user.

If the business connects Google Calendar, Darto shares with Google only the data needed to authenticate the OAuth connection and read from or write to the calendar authorized by the user.

Darto does not use data obtained from Google Workspace APIs to develop, improve, or train generalized AI or machine learning models. Necessary Calendar data may be processed by OpenAI or Google Gemini and directly by TypeSafe AI to evaluate a proposed action. Darto limits this processing to user-authorized appointment features and routes with a no-training commitment; it does not send OAuth tokens to these evaluators. Darto’s use and transfer of information received from Google APIs adhere to the Google API Services User Data Policy, including its Limited Use requirements. No training does not remove other consent, minimization, security, and restricted-access obligations.

If the user requests deletion of their data or disconnects Google Calendar from Darto, Darto deletes the integration link and stops accessing the calendar from that point onward, subject to limited retention that may be required for legal, security, or backup reasons.

10.2 Google Sheets and Google Docs

If the business owner chooses to connect Google Sheets or Google Docs through OAuth, Darto only requests the drive.file permission, which gives access only to the files the user explicitly picks in Google’s file picker, never to the rest of their Google Drive. With that permission, Darto may access and process the following categories of Google data:

  • The email address of the connected Google account.
  • The OAuth permissions granted and the tokens needed to keep the connection active.
  • The ID and name of each selected file.
  • In Google Sheets: the tabs, headers and a sample of rows of each tab during the setup analysis; the content of the tabs the user marks as queryable by their clients; and the rows Darto itself adds or updates for a client, when the user allows it.
  • In Google Docs: the title, the heading structure and the text of the document, or only of the sections the user authorizes.

Darto uses this Google data only to:

  • analyze each file and propose to the user how the assistant can use it, a proposal the user reviews and accepts;
  • answer the business’s clients with information from the tabs or sections the user has authorized;
  • add new rows and update the rows Darto created for a client, only in the tabs where the user allows it and, if the user sets it up that way, after their approval;
  • create, before any change to a sheet’s structure (a new tab or columns accepted by the user), a backup copy in the user’s own Google Drive;
  • detect changes to the structure of a connected file in order to pause its use and notify the user.

Darto never reorders, renames or deletes the existing content of the files, does not edit Google Docs documents and does not access files the user has not selected. Tabs and sections marked as private are never shown to the business’s clients. To answer quickly, Darto keeps a copy of the authorized content, which is refreshed periodically.

Darto does not use data obtained through Google Workspace APIs to develop, improve or train generalized AI or machine learning models. The necessary Sheets and Docs data may be processed by OpenAI or Google Gemini to analyze the files and to answer clients, and by TypeSafe AI directly to evaluate a proposed action. Darto limits this processing to the functions authorized by the user and to routes with a no-training commitment; it does not send OAuth tokens to these providers. Darto’s use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.

If the user removes a file or disconnects Google Sheets or Google Docs in Darto, Darto deletes the tokens, the file configuration and the copy of their content, and stops accessing them from that moment on. Backup copies created in the user’s Google Drive and rows already written to their sheets remain theirs. All of this is without prejudice to any limited retention that may be required for legal, security or backup reasons.

11. Changes to this policy

We will publish any update to this policy at this same URL: https://www.darto.app/privacy