Skip to content
SafetyScorer

Is Wysa safe?

Wysa

Wysa is one of the more careful AI mental health chatbots. You can use it with just a nickname, Apple's label shows no tracking, and one button deletes your history. The catch is that chats go to outside AI models such as OpenAI's after names are stripped, and Wysa says it is not for crises or severe illness.

Medium confidence#10 of 62 in Health, family and learning
Updated
Spot a mistake?

#10of 62 in Health, family and learning

Medium confidence
Agree with our score?
Maker
Touchkin eServices Pvt. Ltd. (India), trading as Wysa
Sign-up
Not needed; a nickname is enough
App Store label
Identifiers, usage data and diagnostics, not linked to you; no tracking
AI models used
Wysa's own AI plus OpenAI, Anthropic and Google Gemini models
Certifications
ISO 27001, ISO 27701, DCB 0129 clinical safety (company statement)
FDA status
Breakthrough Device Designation (May 2022) for one use; not FDA-cleared
Minimum age
13 in the terms; App Store rating 9+
Price (US)
Free to download; Premium $99.99 a year; Premium Plus with a coach $99.99 a month

How the score breaks down

The overall score is the weighted average of these factors. Scoring rules

  • Privacy30% of score

    72 out of 100, Good

    Apple's label lists only identifiers, usage data and diagnostics, none linked to you or used to track you. Chats go to third-party AI models (OpenAI, Anthropic, Google Gemini) after personal details are removed, with zero data retention there. Data can be kept up to 10 years unless you reset it.

  • Security30% of score

    68 out of 100, Fair

    Wysa says data is encrypted in transit and at rest, that it is certified to ISO 27001 and ISO 27701, and that outside auditors check it every year. We found no reported breach. There is no bug bounty that we could find, and chats are not End-to-end encryptionEncryption in which only the sender and the recipient can read a message. The company running the service cannot see its content.More in the glossary.

  • Transparency and control15% of score

    76 out of 100, Good

    You need no account, and "Reset my data" in the app's settings deletes your ID, chats, assessment answers and settings at once. The October 2026 privacy policy is long but written in plain language. Data held by a human coach is not deleted by the reset.

  • Scams and harmful content15% of score

    64 out of 100, Fair

    No contact with other users, and Wysa says safety rules filter every message to and from the AI. Its terms say it is not for crises or for people with severe and enduring mental illness, and admit the AI can give wrong answers. The terms set a minimum age of 13; the App Store rates it 9+.

  • Regulatory record10% of score

    87 out of 100, Excellent

    We found no fine, order or regulator investigation naming Wysa. Its 2022 FDA Breakthrough Device Designation is a development programme, not a review of the consumer app, so the score sits in the band for apps no large regulator has examined.

Is Wysa safe?

For everyday stress, low mood or sleep problems, Wysa is one of the safer AI chatbots to talk to. It does not ask for an email address or real name, Apple's privacy label shows no tracking at all, and a single setting wipes your history. The trade-off is that your messages, with names and other identifiers stripped out, are sent to outside large language models from OpenAI, Anthropic or Google to write the replies. Wysa also says plainly that it is not for crises or for people with severe and enduring mental illness.

We score Wysa 72/100, ahead of Headspace at 69 and far ahead of companion bots such as Replika at 43, which keep intimate chats tied to an account and track users for ads.

What happens to what you tell Wysa

Wysa was built so you can stay anonymous. You choose a nickname, and the app stores data against masked identifiers. Apple's label lists only identifiers, usage data and diagnostics, and none of it as linked to you. Google Play is a little broader: it lists in-app messages and optional health information as collected, and app interactions and device IDs as shared with analytics providers.

Since Wysa added generative AI, the conversation space runs on its own models plus third-party ones by default. The service provider list, updated in August 2026, names OpenAI, Anthropic and Google Gemini. Before a message goes out, a tool from a company called Private AI checks for personal details you may have typed by mistake, and the app asks you to change the message. Wysa says the AI providers keep no conversation data, process it in their European regions and do not train their own models on it. Wysa itself may use anonymous chats to improve its AI.

On advertising, the policy says Wysa does not sell or share your information, messages or app use with advertisers or data brokers. It does use Meta Ads Manager to run ads for the service and Google Analytics with pseudonymous event data.

The weak spot is retention. Unless a contract sets a shorter period, Wysa keeps conversation data, with identifiers removed, for up to 10 years from your last update. If you stop using it, reset your data first.

Is Wysa safe for a mental health crisis?

No, and Wysa does not claim it is. Its terms, revised in June 2026, say the app is for emotional wellbeing support, does not diagnose or give medical advice, is not recommended in crisis situations and is not suitable for people with severe and enduring mental health problems. The privacy policy adds that AI is a developing technology and "you may sometimes get incorrect responses".

Wysa says every message to and from the language model passes through safety filters, and messages that fail them never reach the model. In employer, school or health-service versions, an SOS, self-harm or abuse alert can trigger an email to the organisation with your name and the message.

The age rules do not line up. The terms set a minimum age of 13, or 11 through a school or health service, and say teenagers get a restricted experience. The App Store rates the app 9+, which is lower than the terms allow.

What the FDA designation means

In May 2022 Wysa announced that the US Food and Drug Administration had granted it a Breakthrough Device Designation for a conversational CBT tool for adults with chronic musculoskeletal pain and depression and anxiety. The designation gives a company faster access to FDA staff while it develops a product. It is not a clearance or approval. Wysa's own note on the announcement says it does not cover all Wysa products, and we found no 510(k) clearance or device registration for Wysa or its maker in the FDA's openFDA data. The consumer app is a wellbeing tool, not a regulated treatment.

How to use Wysa more safely

  1. Keep the nickname vague and leave out names, places and other people's details, even though Wysa tries to strip them.
  2. Use "Reset my data" in settings when you stop, or your data may be kept for years.
  3. Save your local crisis number in your phone. In the US, call or text 988; in the UK, call 111 or Samaritans on 116 123.
  4. Check what a work, school or NHS version shares with the organisation before you use it, because alerts there include your name.
  5. Cancel Premium through the App Store or Google Play: Wysa gives no refunds for unused time.
  6. If what you discuss is getting heavier, talk to your doctor about care from a licensed professional, such as Talkspace or a local service.

How we scored it

Regulatory record (87) is Wysa's strongest factor, because no regulator has fined or ordered it. Transparency and control (76) and privacy (72) come next: no account, no tracking on Apple's label and a one-tap reset, held back by outside AI models and up to 10 years of retention. Scams and harmful content (64) is the weakest, because an AI is giving mental health support to people who may be vulnerable, and Wysa itself says it is not for crises. Security (68) rests on Wysa's ISO certifications and a clean breach record, without a bug bounty.

Questions people ask

Is Wysa a real therapist?

No. The free app is an AI chatbot that walks you through exercises based on CBT, DBT and mindfulness. A paid plan adds a human coach, but Wysa's terms say the service does not give medical advice or diagnoses and is not a substitute for a professional. Talk to your doctor if you think you need treatment.

Is Wysa approved by the FDA?

No. In May 2022 the FDA gave Wysa a Breakthrough Device Designation for an AI-led CBT tool for adults with chronic musculoskeletal pain plus depression and anxiety. That designation speeds up talks with the agency before a review; it is not clearance or approval, and Wysa itself says it does not cover all its products.

Can anyone read my Wysa chats?

Wysa says no real person can see what you write in the chatbot. Your messages are sent to outside AI models from OpenAI, Anthropic or Google to generate replies, after a separate service strips personal details, and Wysa says those providers keep nothing and do not train on them. If you use a human coach or an employer or NHS version, that coach or clinician may see your conversations.

What happens if I tell Wysa I am in danger?

Wysa says its safety rules check messages for risk, and in employer, school or health-service versions an alert can go to that organisation. Its terms are clear that it is not for emergencies or crises. If you are in danger, call your local emergency number; in the US, call or text 988.

How do I delete my Wysa data?

Open the app's settings and choose "Reset my data". It deletes your ID, past chats, reminders, assessment answers and settings, and cannot be undone. Data kept by a coach or medical assistant is not removed this way, so ask Wysa's privacy team by email for that.

Sources

  1. 1.Wysa Global Privacy Policy (version 7.2.2, revised 16 October 2026) (opens in a new tab) · Wysa, 2026
  2. 2.Wysa Global List of Service Providers (revised 26 August 2026) (opens in a new tab) · Wysa, 2026
  3. 3.Wysa Terms of Service (revised 5 June 2026) (opens in a new tab) · Wysa, 2026
  4. 4.Wysa: Mental Wellbeing AI, App Store listing and App Privacy label (opens in a new tab) · Apple App Store, 2026
  5. 5.Wysa, Google Play data safety (opens in a new tab) · Google Play, 2026
  6. 6.Wysa Receives FDA Breakthrough Device Designation for AI-led Mental Health Conversational Agent (opens in a new tab) · Wysa, 2022
  7. 7.openFDA device 510(k) and registration search (no Wysa or Touchkin records) (opens in a new tab) · U.S. Food and Drug Administration, 2026
Spot a mistake?