Legal

Westercove™ AI Use Policy

Last updated: November 2, 2026. Version 3.

Part of the Westercove Terms & Conditions.

The short version

Westercove's companion is artificial intelligence made by Anthropic. It sees what you write so that it can respond, and it is never trained on your writing. It can be wrong. It will never make a decision about your care, and when a message suggests danger, Westercove shows you people who can help.

How Westercove™ uses AI

Westercove uses AI models from Anthropic, accessed through Anthropic's commercial API (with Amazon Bedrock as a backup), for four things: the companion's responses; summaries of your writing; the Grief Library feature, which offers reflections and practices drawn from resources you have chosen; and a safety check that reads what you write in context to decide whether to show you support or crisis resources. Our access keys are held on our servers and never on your device. If you use voice typing, your device’s or browser’s speech service turns your speech into text before anything is sent. Westercove does not offer spoken conversations with the companion: it always replies in writing.

What is sent to the AI

The message you just wrote, the recent conversation, and the context the feature needs. That context may include the name you use for the person you are grieving, relevant earlier entries, your preferences (for example, whether faith language is welcome), and your Grief Library choices.

What is never sent

Your password, your payment details, your email address, and anything that belongs to another Westercove user. We send only what the feature you are using needs.

Training and retention

Anthropic processes what we send only to return a response to us. Under the commercial terms we hold with Anthropic, your content is not used to train its models. If we ever change providers, or those terms change, we will update this policy first.

How to read what the AI says

Treat every response as a reflection to consider, never as a verdict. AI can state something false with confidence. It knows only what you have told it. It cannot see your whole life, and it is not a clinician. It is not a medical device and is not intended to diagnose or treat any condition.

Safety measures, and their limits

The safety check reads the whole entry, not single words, so that writing about a loss is not mistaken for a crisis. Direct statements that you intend to end your life show crisis resources immediately, even if the AI is unavailable. These checks are fully automated. At the highest level of concern in companion chat, our team receives an alert with the type of concern and an internal reference number, never your name or what you wrote. Your entry is still saved, and you are shown crisis resources right away. Westercove will not contact anyone on your behalf. These measures reduce risk. They do not remove it. No software catches everything, which is why 988, 741741, and 911 are always on your screen.

Honesty about AI

Westercove will always tell you that you are talking with AI. It will never claim to be a person, a therapist, or a clinician. We use Anthropic's services in line with its usage policies, including those for consumer-facing uses involving health and wellbeing. The companion is not a licensed mental health professional, counselor, or therapist, and it cannot act as one. If you have a concern about something the companion said, write to info@westercove.com.

On every journaling screen, directly under the space where you write your entry, Westercove shows: “Your AI grief support companion offers grief education, guided journaling, and personalized support. It will respond to your entry, and you can continue the conversation for as long as you need.”

Back to the Westercove™ home page
In crisis? Call or text 988. Text HOME to 741741.