Legal

Privacy policy

Last updated: 20 August 2026

Who we are

Miral builds AI customer support agents. Our clients connect their support channels and their knowledge to Miral, and the agent answers their customers on those channels. This policy covers what we collect to do that, how it is used, who processes it on our behalf, and how to have it deleted. It applies to this website, to the Miral dashboard at app.meetmiral.com, and to conversations handled by a Miral agent.

Three kinds of people interact with Miral: clients, the companies that hire an agent; end users, the customers and staff who talk to it; and visitors to this website. What we hold about each is different, so this policy treats them separately.

What we collect

  • Conversations. Messages sent to a Miral agent, on the chat widget, in Slack, by email, or as a comment under a client's Facebook or Instagram post, along with the replies. If you share an email address or other details inside a conversation, they are part of that conversation. For a comment, we hold the comment's text, the public name on the account that wrote it, and the identifiers Meta assigns to the comment and the post.
  • Connected accounts. When a client connects a Facebook Page and its Instagram account, we hold the access tokens Meta issues, the Page, Instagram account and ad account identifiers and names, and the identifier of the Facebook account that granted them. We use them only to read and answer comments on that Page and, where the client asks, to build advertising campaigns in that account for the client to review.
  • Client knowledge. Documents our clients connect so their agent can answer from them: help center articles, policies, product documentation, internal notes. This content belongs to the client and stays theirs.
  • Account data. For dashboard users: your email address, the sign-in codes we send to it, and the settings you save.
  • Service logs. Technical records of the service operating: timestamps, message identifiers, errors. We use these to keep the service reliable and to investigate problems.

This website itself sets no analytics or advertising cookies and does not track visitors. The chat widget on the homepage starts a conversation only when you open it and write something.

How we use it

To answer. A message sent to a Miral agent is read together with the relevant parts of that client's knowledge, and a reply is generated. When the agent hands a conversation to a human, that person sees the thread the agent saw. We also use conversation history so the agent remembers context within a conversation, and service logs to operate and improve the reliability of the service.

We do not sell your data. We do not use your conversations or your documents to train AI models. We do not use them for advertising.

One client, one boundary

Each client's data is kept separate. One client's documents, conversations, and settings are never used to answer another client's customers.

Who processes data on our behalf

We use a small number of service providers to operate Miral: cloud hosting, AI model providers that generate replies, email delivery for things like sign-in codes, and the messaging platforms a client connects, such as Slack, WhatsApp, and Meta's Facebook and Instagram, which are governed by their own terms. Each provider receives only what its role requires, and the current list is in our data processing agreement.

We share data with no one else, except where the law requires it.

Retention and deletion

Conversations and connected documents are kept while the client's account is active, so the agent can use history and the client can review what it said. A client can download a copy of their data and delete their workspace themselves from Settings, which cancels their plan and removes their conversations, documents, seats and connected credentials at once; what remains in backups goes within 30 days.

To have your data corrected or deleted sooner, write to us at the address below. If you spoke with a Miral agent as a customer of one of our clients, that client controls the conversation, and we will route your request to them or act on their instruction.

Facebook and Instagram. If you connected a Facebook Page to Miral and want that connection and its data removed, disconnect Meta on the Integrations screen of the dashboard, or remove Miral under Settings, Business integrations in your Facebook account; either one deletes the tokens and identifiers we hold at once. Facebook's own "Delete data" request for Miral reaches us the same way and is carried out immediately, with a confirmation code you can check. If you left a comment under a client's post and want it removed from Miral, write to us or to that client; the comment itself stays on Facebook or Instagram unless you or the Page delete it there.

Security

All data moves over encrypted connections. Credentials for connected channels are stored encrypted. Access to production systems is limited to the people who operate the service.

No one can honestly promise perfect security, so we will tell affected clients promptly if an incident touches their data.

Changes and contact

When this policy changes, the date at the top changes with it, and we will tell active clients about anything material. Questions and requests: support@meetmiral.com.