Myndchat Privacy Policy
Myndchat is a conversation automation platform. Businesses use Myndchat to serve their own customers over WhatsApp, Instagram Direct, Facebook Messenger and TikTok. This document explains what we do with personal data: yours, if you subscribe to Myndchat or visit this website, and other people's, when it passes through the platform on behalf of one of our customers.
We process personal data in two situations that the law keeps apart, and that this document keeps apart with it: on our own decision, when you subscribe to Myndchat or browse this website, and on a customer's instruction, when it is that customer who decides what happens to the data of the people talking to them. Section 2 explains the difference and sits at the very beginning on purpose, because it determines whom you should ask for what.
This policy describes the processing of personal data of the following data subjects:
- people who subscribe to Myndchat and the members of that organisation's team, including those who only operate the platform day to day;
- visitors to myndchat.com;
- anyone who writes to us at any of the addresses published in section 1;
- the end contacts who talk to our customers over the connected channels — for those, Myndchat is a processor and the customer is the one answerable for the processing, as section 4 explains.
It covers the myndchat.com website and the Myndchat platform, including the integrations with WhatsApp Business Platform, Instagram, Messenger and TikTok. It does not cover what each customer does on its own outside Myndchat, nor what the messaging platforms do with your data as owners of the channel: each has its own policy, and section 9 identifies all of them.
If you have a question about any point in this document, or want to exercise a right under the Brazilian General Data Protection Law, write to privacidade@myndchat.com. Section 13 states how long we take to answer and how we confirm the identity of whoever asks before handing over personal data.
Table of Contents
- 1. Who we are and how to reach us
- 2. The distinction that organises this policy
- 3. Data we process as controller
- 4. Data we process as processor
- 5. Isolation between customers
- 6. Cookies and similar technologies
- 7. Artificial intelligence
- 8. Automated decisions and the right to review
- 9. Sharing and subprocessors
- 10. International transfers
- 11. Information security
- 12. Retention and deletion
- 13. Your rights as a data subject
- 14. Security incidents
- 15. Children and adolescents
- 16. Application access logs
- 17. Changes to this policy
- 18. Effective date and version
1. Who we are and how to reach us
Myndchat is a conversation automation platform that businesses use to serve their own customers over WhatsApp, Instagram Direct, Facebook Messenger and TikTok. This policy explains what we do with personal data, in which situations, on what legal basis and for how long.
| Field | Value |
|---|---|
| Legal name | MYNDCHAT TECNOLOGIA LTDA |
| Company registration (CNPJ) | 65.434.606/0001-80 |
| Address | Avenida Brigadeiro Faria Lima, 1811, Sala ESC 1119, Jardim Paulistano, São Paulo/SP, 01452-001, Brazil |
| Phone | +55 48 98434-0133 |
| Data Protection Officer | Bernardo Prantz Pin |
| Data Protection Officer contact | encarregado@myndchat.com |
| Privacy and data subject rights | privacidade@myndchat.com |
| Support | support@myndchat.com |
| Security and incidents | seguranca@myndchat.com |
| Report abuse | abuso@myndchat.com |
The Data Protection Officer is the point of contact between you, us and the Brazilian data protection authority. You may write to them directly about anything in this policy, including complaints.
About the moment this policy takes effect. As of this date, the Myndchat website is live and the platform is under construction. The processing described in sections 3, 4, 5, 7, 8 and 12 concerns the platform and begins, for each customer, when they create an account and connect a channel. Until then, the only processing under way is the one described in section 6, which covers the website. We publish the policy in advance because whoever signs up must be able to read what will be done with the data before deciding, and because the platforms we integrate with require the policy to be published and live before authorising the integration.
2. The distinction that organises this policy
This is the most important section of the document, and it sits near the top on purpose. Myndchat holds two different roles at the same time, and what you can require from us depends on which one applies to you.
We are the CONTROLLER of the data of whoever subscribes to the platform. If you signed up for Myndchat, or you are part of the team of a business that did, we decide why and how we process your account, billing and product usage data. In that case, you exercise your rights directly with us.
We are the PROCESSOR of the data of our customers' end contacts. If you spoke to a shop, a clinic or an agency over WhatsApp, Instagram, Messenger or TikTok, and that business uses Myndchat, then the conversation passes through our systems — but the party that decides the purpose of that processing, collects your consent and answers for it is the business you spoke to, not Myndchat. We process that data under their instructions.
If you are an end contact and want to access, correct or delete your data, the route is to contact the business you spoke to. They are the controller. If you write to us anyway, we will not ignore you: we forward the request to the responsible customer, confirm to you that we did, and support the response within what is technically available to us. What we cannot do is decide, on their behalf, about data that is not ours.
This separation is not legal formality: it changes what we do in practice. No data processed in our role as processor is used for our own purposes, cross-referenced between different organisations, sold, shared for advertising or used to train artificial intelligence models. Section 4 sets that out.
3. Data we process as controller
This is the data of whoever subscribes to Myndchat and of that organisation's team. Below is what we collect, what for, on what legal basis and for how long we keep it.
| Data | What it is for | Legal basis | How long |
|---|---|---|---|
| Name, e-mail and phone | Identify the account holder, authenticate access and communicate about the service | Performance of a contract (art. 7, V) | While the account is active, plus 5 years |
| Password, in hashed form | Authenticate access | Performance of a contract (art. 7, V) | While the account exists |
| Second authentication factor and recovery codes | Protect the account against unauthorised access | Performance of a contract and legitimate interest in security (art. 7, V and IX) | While the feature is enabled |
| Tax identification number of the responsible party | Issue tax documents, identify the contracting party and prevent fraud | Legal obligation and performance of a contract (art. 7, II and V) | 5 years after the last tax document |
| Billing address | Issue tax documents and charge the subscription | Legal obligation and performance of a contract (art. 7, II and V) | 5 years |
| Subscription, plan, add-ons and payment status | Grant the contracted access and charge correctly | Performance of a contract (art. 7, V) | 5 years |
| Billing identifiers, card brand, last four digits and expiry date | Display the stored payment method and process the recurring charge | Performance of a contract (art. 7, V) | While active, plus 5 years from the last charge |
| Record of acceptance of the terms, with version, date, time and IP address | Prove that the contract was accepted and in which version | Performance of a contract and exercise of rights (art. 7, V and VI) | While the account is active, plus 5 years |
| Audit trail: who did what, when and from which IP address | Security, incident investigation and due process | Legitimate interest and exercise of rights (art. 7, IX and VI) | 24 months |
| Application access logs | Comply with art. 15 of the Brazilian Internet Civil Framework | Legal obligation (art. 7, II) | 6 months |
| Error reports, with route, screen context and screenshot | Diagnose a defect you reported | Legitimate interest (art. 7, IX) | 90 days |
| Usage metrics, active contact count and artificial intelligence consumption | Apply plan limits, warn before blocking and charge overage | Performance of a contract (art. 7, V) | 24 months, and 5 years for what supported an invoice |
| Access credentials for your channel accounts | Send and receive messages on your behalf, which is the core function of the service | Performance of a contract (art. 7, V) | Until you disconnect the channel or close the account |
| IP address and request data | Limit abuse, prevent fraud and protect the platform | Legitimate interest (art. 7, IX) | 6 months |
We never collect the full card number or the security code. Those fields are filled in inside an isolated component provided and operated by our payment processor, and they do not pass through our servers, our logs or our backups.
4. Data we process as processor
This is the data of the people who talk to our customers. Here Myndchat decides nothing about purpose: the controller is the business using the platform, and it defines why it talks to you, on what legal basis and for how long it intends to keep the data.
| Data | Where it comes from | Default retention |
|---|---|---|
| Contact name and channel identifier | From the messaging platform itself, when the conversation arrives | 12 months |
| Phone number, in the case of WhatsApp | From the messaging platform | 12 months |
| Profile picture, when the channel sends it | From the messaging platform | 12 months |
| Full message content, in both directions | From the conversation | 12 months, configurable to less by the customer |
| Images, audio, video and documents exchanged | From the conversation | 6 months, configurable to less by the customer |
| Tags, custom fields and segments | Created by the customer or by their automation | 12 months |
| Lead qualification | From the automation or the artificial intelligence layer | 12 months |
| Contact and conversation variables, and contact memory | Written by an explicit action of an automation or an agent | 12 months, and conversation variables are discarded when the conversation ends |
| Deals, sales pipeline and journey events | Recorded by the customer or by their automation | 12 months |
| Contact source and originating advertisement | From the messaging platform | 12 months |
| Participation in campaigns and sequences | From the sends made by the customer | 12 months |
| Record of opt-in and opt-out | From the contact's own statement or the customer's record | For as long as the organisation exists |
| Internal team notes about the contact | Written by the agent, never sent to the contact | 12 months |
The opt-out record has a different retention period on purpose: it survives the deletion of the rest of that contact's data. If we deleted the opt-out along with everything else, the person would start receiving messages again on the next import, and their request would stop being honoured.
What we never do with this data, without exception:
- We do not use it for Myndchat's own purposes.
- We do not cross-reference data between different customer organisations.
- We do not sell it, rent it or transfer it to third parties.
- We do not use it for advertising, ours or anyone else's.
- We do not use it to train artificial intelligence models, ours or third parties'.
- We do not access conversation content outside the cases in section 11, and every internal access is logged and visible to the customer in their own audit trail.
This is not only our own policy. Our contract with Meta to operate as a technology provider requires us to process data obtained through its platforms exclusively on behalf of the customer and according to their instructions, and forbids using it for any other purpose or disclosing it to third parties.
5. Isolation between customers
The platform is used by several organisations at the same time on the same infrastructure. What stops one from seeing another is not the way the screens are arranged: it is the architecture.
- Each organisation is a data boundary in the database, enforced by the database itself and not only by the application. A query attempting to reach another organisation's data is rejected at the source.
- The check also exists on the server, independently. The two layers work together: if one fails through a mistake of ours, the other keeps blocking.
- The isolation applies equally to the store used by the artificial intelligence similarity search. One customer's knowledge base does not reach another.
- The identifiers used in page addresses are not sequential, which prevents discovering third-party records by guessing.
- Files and media live in a path separated by organisation and are served only through a temporary, signed address. There is no permanent public link to conversation media.
- Real-time access to the inbox is authorised per organisation. Attempting to subscribe to another organisation's channel is blocked and logged.
We test this isolation by attacking it: we create separate organisations and try, through every available route, to read, alter and delete one organisation's data from the other. A control is only considered valid when there is a test that fails if the control is removed.
6. Cookies and similar technologies
This section summarises what we do with cookies. The full detail — the measured inventory, what each category does and how to block cookies through your browser — is in the Cookie Policy, which is a document of its own and is listed alongside. Where the two overlap, the Cookie Policy prevails, being the more specific one about your browser.
A cookie is a small file that a website stores in your browser. We apply the same treatment here to equivalent technologies, such as browser local storage and measurement pixels.
Nothing that is not strictly necessary loads before you decide. Until you choose, no third-party script runs, no pixel fires and no non-essential cookie is stored. Closing the notice, clicking outside it or continuing to browse is not consent: without an affirmative action from you, nothing is released and the notice comes back.
We use four categories, and they are the same in the notice, in the settings panel and here:
| Category | What it is for | Initial state | Legal basis |
|---|---|---|---|
| Essential | Making the website and the platform work: keeping the session, remembering the chosen language, protecting forms against forged requests and storing your own consent choice | Always on, and cannot be turned off | Legitimate interest, as strictly necessary (art. 7, IX) |
| Preference | Remembering interface choices that are not required for the service to work | Off | Consent (art. 7, I) |
| Analytics | Measuring audience and understanding how the site is used, in order to improve it | Off | Consent (art. 7, I) |
| Advertising | Displaying and measuring targeted advertising | Off | Consent (art. 7, I) |
The cookies and technologies actually in use today, measured in the browser rather than estimated:
| Name | Category | Who stores it | What it holds | Duration |
|---|---|---|---|---|
| Language cookie | Essential | Myndchat (first party) | Only the chosen language code, pt or en. It holds no identifier and cannot be used to recognise you | 1 year |
| Consent cookie | Essential | Myndchat (first party) | Your choice per category, when it was made and the version of this policy | 6 months, after which the question is asked again |
We use no service in the Preference, Advertising and Analytics categories at this time — all three are empty, and we measured that in the browser rather than assuming it. The categories still exist in the panel because, if a service ever enters any of them, that service starts off and depends on your consent — rather than on a silent change to the structure of the notice.
You can change your mind at any time, free of charge and in the same place: the Privacy Settings link, in the footer of every page, reopens the panel. Withdrawing is as simple as accepting, and withdrawal actually removes the cookies in that category rather than merely recording the refusal. You can also block or delete cookies through your browser settings; in that case some functions may stop working.
We keep a record of your consent, as the law requires of anyone relying on it: the consent identifier, what was accepted and what was refused, when, the version of this policy and the origin of the choice. The identifier is shown to you inside the Privacy Settings panel, so that you can ask us for the data of your own consent.
We ask the question again when this policy changes version, when a new service enters a non-essential category, and every six months.
7. Artificial intelligence
Myndchat uses artificial intelligence in tiers, and most conversations never reach a generative model. We describe here exactly what leaves our infrastructure, where it goes and under what guarantee.
The automation resolves in layers, and only moves to the next when the current one cannot answer:
| Layer | What it does | What leaves our infrastructure |
|---|---|---|
| Flowchart | The decision tree built by the customer themselves. It always tries first | Nothing. No content leaves |
| Similarity search | Looks for the answer in a question-and-answer base registered by the customer, comparing meaning rather than exact wording | The text is converted into a numeric representation by an external provider. No text generation happens in this layer |
| Generative model | Last resort: it only runs when neither the flowchart nor the similarity search resolved the message | A slice of the conversation, described below |
When the last layer runs, what we send to the provider is bounded before the call and capped: at most the last fifteen messages of that conversation, plus a compacted summary of what came before, plus the business information the customer registered, plus the variables and memory of that contact which are relevant to answering.
What is never sent to an artificial intelligence provider:
- No password, key, credential or secret.
- No data from another organisation, under any circumstance — including in the similarity search, which is scoped to the organisation in the query itself and not only by the database rule.
- No billing data and no identity document.
- No personal data beyond what is necessary to answer the message at hand.
The model also has no power to act on its own. It produces a suggested reply and a classification; any action with real effect only happens if the output matches a closed list of permitted actions, checked by our own code. The model does not write to the database, does not change permissions, does not touch billing and does not decide account suspensions.
About the providers that receive this content, and under what guarantee:
| Provider | Function | Use for model training | Retention by the provider |
|---|---|---|---|
| Anthropic | Generating replies, conversation summaries and translation | Prohibited by the commercial application programming interface terms applicable to our account | Deletion within 30 days. Content flagged for violating the provider's usage policy may be retained longer by them |
| OpenAI | Converting text into a numeric representation for the similarity search | Data sent through the application programming interface is not used to train models, absent an express opt-in that we have not made | Abuse monitoring logs for up to 30 days |
We describe what is actually contracted, not what would be nicer to announce. The training prohibition is contractual and in force. The short abuse-monitoring retention window exists and is real: we have requested a zero data retention agreement from both providers, which is subject to their prior approval. Until such an agreement is in force, what is written in the table above applies, and this policy will be updated, with a new version, on the day that changes.
We log, per call, which category of data was sent. That is what allows us to answer a data subject request about what was shared with precision.
8. Automated decisions and the right to review
Part of what the platform does is decided automatically, without human intervention at the time. You have the right to know what that is, on what criteria, and to request a review.
| What is decided automatically | On what criteria | What it means for you |
|---|---|---|
| Qualifying a contact as a lead | The content of the conversation and the rules the business itself registered about its own operation | It changes the priority and the order in which the business talks to you. It does not deny service, does not set prices and does not affect credit |
| Routing the conversation to a person or a team | Rules configured by the business over data it already holds: intent marked in the conversation, tag, channel, time of day. No model is called for this decision | It defines who on the team serves you |
| Automatic reply by the flowchart or by artificial intelligence | The tree built by the business and the business information it registered | You receive an automated reply, with escalation to a person when you ask, when the subject falls outside scope, or when confidence is low |
None of these decisions restricts your access to a product or service, sets a price for you, affects a credit assessment or produces legal effects. Even so, the right to review applies.
To request a review of an automated decision, write to privacidade@myndchat.com. If the decision was made inside a customer's account, we forward the request to them as the controller and let you know. The review is carried out by a person. We respond within 15 days.
On our side, as controller, we also use automated processing to detect abusive use of the platform by our own customers — unsolicited messaging, scams, prohibited content. Warnings and observation may be automatic. Limiting or suspending an account requires a decision by a person, recorded with author and reason, and the customer has an appeal channel with human review. Artificial intelligence may flag and prioritise; it never decides a suspension.
9. Sharing and subprocessors
We do not sell personal data and we do not share it for third-party advertising. We share only with suppliers that carry out part of the service on our behalf, and only what is necessary for that function. Each of them is bound by contractual obligations of confidentiality, security and purpose limitation.
The public, versioned and complete list is on the Subprocessors page, which is the official source: there each provider appears separated by state — who already processes data today, who is planned for the platform, and which functions have no provider chosen yet. The summary below repeats the planned providers, without that separation:
| Supplier | Function in the product | Category of data received | Where it processes |
|---|---|---|---|
| Supabase | Database, authentication, media storage, real time and similarity search | Practically all data in sections 3 and 4 | Outside Brazil |
| Vercel | Hosting, page delivery and scheduled task execution | Request data, IP address and content in transit | Outside Brazil |
| Stripe | Subscription billing and payment processing | Name, e-mail, tax identification number, billing address and card data, the latter collected directly by their component | Outside Brazil |
| Meta Platforms | WhatsApp, Instagram Direct and Facebook Messenger channels | Message content, phone number, contact identifier and profile, media | Outside Brazil |
| TikTok | TikTok direct message channel | Message content and contact identifier | Outside Brazil |
| Anthropic | Reply generation, summary and translation in the artificial intelligence layer | The conversation slice described in section 7 | Outside Brazil |
| OpenAI | Converting text into a numeric representation for similarity search | Text of the registered questions and answers and of the indexed conversations | Outside Brazil |
| Assisted address completion in forms | The text typed into the address field and the IP address | Outside Brazil | |
| Google Analytics | Audience measurement for the website | Measurement identifier and browsing data, only with your consent | Outside Brazil |
| Sentry | Error capture with already filtered context | Route, correlation identifier and a fragment of screen state | Outside Brazil |
Almost none of them processes data yet. As of this date only the website is live, and the only provider actually processing personal data is the one hosting and delivering the pages. The others begin processing when the corresponding platform function goes into operation. The Subprocessors page states, provider by provider, which state each one is in.
We choose suppliers on technical capability, security maturity and willingness to take on by contract the obligations the law imposes on us. Before adding a new supplier to this list, we assess those three things. A change of supplier is announced to customers at least 30 days in advance, produces a new version of the Subprocessors page and a new version of this policy.
We may also share data where there is a court order, a request from a competent authority under the law, or a need to exercise our rights in proceedings. In those cases we inform the data subject whenever the law allows.
10. International transfers
All the suppliers listed in section 9 process data outside Brazil. That is an international transfer, and it requires its own legal ground.
| Destination | Ground for the transfer |
|---|---|
| European Union, European Economic Area and European Union institutions | Adequacy decision issued by the Brazilian data protection authority, recognising those destinations as providing an adequate level of protection (art. 33, I of Law 13.709/2018) |
| United States and other destinations | Standard contractual clauses approved by the Brazilian data protection authority, adopted in full and without modification, under art. 33, II, item b of Law 13.709/2018 |
You have the right to receive the full text of the clauses used in a transfer involving your data. Ask at privacidade@myndchat.com and we respond within 15 days, preserving third-party trade secrets where applicable.
Brazilian law continues to apply to processing carried out abroad whenever the collection takes place in Brazilian territory or the service is offered to the Brazilian public.
11. Information security
We hold private conversations belonging to people who never signed up for Myndchat. That demands above-average rigour, and the measures below are the ones we actually apply.
- Encrypted communication in transit across the entire service, without exception.
- Access credentials for your channel accounts encrypted at rest, with a key managed outside the database, so that a copy of the database does not hand them over. They are never sent to the browser, never appear in logs and are never displayed, not even partially.
- Passwords checked against public breach databases at sign-up and at password change, rejecting those already known to have leaked.
- Second authentication factor available to everyone and required for the administrator role on paid plans.
- Least privilege access: each person and each component sees only what their own function requires.
- Isolation between organisations enforced in the database and on the server, independently, as described in section 5.
- An immutable audit trail recording who accessed what, when and from where.
- Internal access to customer data is exceptional, requires a written justification, is time-limited and is fully logged. Access to conversation content appears in the organisation's own audit trail, visible to the customer.
- Automated scanning for credentials in the code and for vulnerabilities in dependencies, on every change.
- Automated tests that attempt to break isolation between organisations and that fail if the control is removed.
- Backups from the first record onward, with a restore procedure that is tested rather than merely documented.
- Sensitive data kept out of error and diagnostic logs, by listing what is allowed rather than what is forbidden — so a new field starts outside the log.
We do not claim certifications we do not hold. None of the measures above is presented as a seal, a standard or a third-party audit: they are controls we implement and test. If and when we obtain a certification, it will be stated here by name and date.
If you find a security flaw, write to seguranca@myndchat.com. We review every report received and we take no action against anyone who reports in good faith, without exploiting the flaw beyond what is needed to demonstrate it and without accessing third-party data.
12. Retention and deletion
We keep each type of data for as long as the purpose justifying it requires, or for the period the law demands. When the period ends, the data is deleted — and deletion here means removal from the database, from file storage and, in the due cycle, from backups as well.
| Data | Period | Counted from |
|---|---|---|
| Application access logs | 6 months | Date of the record |
| Conversations and messages | 12 months, configurable to less by the customer | Last message in the conversation |
| Images, audio, video and documents | 6 months, configurable to less by the customer | Date received |
| Automation execution traces and internal events | 90 days | Date of the event |
| Audit trail | 24 months | Date of the record |
| Record of acceptance of the terms | While the account is active, plus 5 years | Account closure |
| Billing and tax data | 5 years | Issue of the document |
| Security incident records | 5 years, at minimum | Date of the record |
| Error reports and screenshots | 90 days | Date of the report |
| Cookie consent records | While valid, plus 5 years | Withdrawal or expiry |
| End contact opt-out records | For as long as the organisation exists | — |
| Channel credentials | Revoked immediately | Channel disconnection or account closure |
When an account is closed, the sequence is this:
- You have 30 days to export your data in a readable format.
- After that period, the data is deleted from the database and from storage within 60 days.
- Backups still containing that data are overwritten in the normal cycle, within 90 days.
- Credentials for connected channels are revoked immediately on closure, not at the end of the periods above.
A period may be suspended, for that specific item only, where there is a longer legal retention obligation, a court order or a need to defend ongoing proceedings. In that case the data is blocked for any other purpose.
13. Your rights as a data subject
Brazilian law grants you the rights below. They apply whether you are our customer or an end contact of a customer — what changes is the route, explained at the end of this section.
| Right | What it means in practice |
|---|---|
| Confirmation | Knowing whether we process any data about you |
| Access | Receiving a copy of the data we hold about you |
| Correction | Correcting incomplete, inaccurate or outdated data |
| Anonymisation, blocking or deletion | Requesting removal of unnecessary or excessive data, or data processed outside the law |
| Portability | Receiving your data in a format that lets you take it to another provider |
| Deletion of data processed on consent | Where the basis is your consent, requesting that the data be erased |
| Information about sharing | Knowing which public and private entities we share your data with |
| Information about refusing | Knowing that you may withhold consent and what happens if you do |
| Withdrawal of consent | Withdrawing, at any time and free of charge, a consent you have given |
| Review of automated decisions | Asking for a person to review a decision taken solely by automated processing, as described in section 8 |
| Objection | Objecting to processing that relies on a legal basis other than consent |
How to exercise them, and what to expect:
- Write to privacidade@myndchat.com, or directly to the Data Protection Officer at encarregado@myndchat.com.
- We respond within 15 days. If the request is complex and we need more time, we tell you within that period, with the reason.
- Before acting, we confirm the request really is yours. Verification uses data we already hold and is proportionate to what is being asked; for access and deletion requests it is stricter, because answering the wrong request is itself a data breach.
- There is no charge. We do not charge for the exercise of a right.
- If we refuse a request, we say why and on what grounds.
If you are an end contact of one of our customers, the route is different. The party controlling your data is the business you spoke to, and the request should be addressed to them. If you do not know how to reach them, or prefer to talk to us, write anyway: we forward your request to the responsible customer, confirm to you that we have done so, and support the response technically. What we do not do is decide, on our own, about data we do not control.
You may also lodge a complaint with the Brazilian data protection authority. We would rather resolve matters directly with you, but that right exists and does not depend on contacting us first.
14. Security incidents
We have a written incident response procedure, defined before we needed it, with the steps of contain, assess, fix, communicate, record and learn.
Where a security incident may result in relevant risk or harm to data subjects, we notify the Brazilian data protection authority and the affected data subjects within three business days, counted from the moment we know the incident affected personal data. The notice describes the nature and category of the affected data, the number of data subjects, the security measures in place before and after, the risks involved and what we will do to reverse or mitigate the effects.
Notice to data subjects is given in plain language and directly, whenever it is possible to identify who was affected. Where it is not, we publish it prominently for a minimum of three months.
We keep a record of every incident for a minimum of five years, including those that did not require notification, together with the assessment that led to that conclusion.
Where an incident affects data we process as processor, we immediately notify the customer acting as controller and support them in meeting their own obligations, as it is they who notify the authority and the data subjects.
15. Children and adolescents
Myndchat is a work tool contracted by businesses. It is not intended for anyone under 18, is not directed at children or adolescents, and does not knowingly collect their data for our own purposes. Creating an account requires legal capacity and business use.
Conversations handled by our customers may in practice involve children or adolescents, because whoever writes to a shop on WhatsApp can be anyone. In those cases the controller is the customer, and it is their duty to process that data in the best interest of the minor and to obtain specific, prominent consent from at least one parent or legal guardian where the law requires it.
If we learn that data of children or adolescents is being processed on the platform contrary to the law, we notify the responsible customer, require them to put it right, and may limit or suspend the feature involved. If you believe this is happening, write to privacidade@myndchat.com.
16. Application access logs
As an internet application provider incorporated as a legal entity and operating for economic purposes, we keep application access logs — date, time and IP address of each access — for six months, under confidentiality and in a controlled environment, in accordance with art. 15 of the Brazilian Internet Civil Framework.
These logs are only disclosed under a court order, and the content of private communications is only disclosed in the cases and in the manner the law establishes. A competent administrative authority may request registration data as provided by law.
We do not keep logs of access to other internet applications and we do not keep personal data that is excessive in relation to the purpose that justified it.
17. Changes to this policy
This policy may change. When it does, we publish a new version with its own number and effective date, and we keep previous versions reachable through the link at the top of this page.
If the change is material — a new purpose, a new legal basis, a new subprocessor in a non-essential category, a change to a retention period or to an international transfer destination — we notify customers prominently and in advance, rather than merely swapping the text.
A version change to this policy or to the Terms of Service triggers a new acceptance inside the platform. Until it is accepted, the customer continues handling existing conversations, and bulk sending features are temporarily unavailable. A version change also causes the cookie question to be asked again.
18. Effective date and version
| Field | Value |
|---|---|
| Version | 1.0 |
| Effective date | 18 August 2026 |
| Previous version | None. This is the first published version |
| Languages | Brazilian Portuguese and English |
In the event of any discrepancy between the Portuguese and the English version of this document, the Portuguese version prevails. This policy is governed by Brazilian law.