Plain Help collects no information that could identify you.
There are no accounts, no login, no email address, no name. Your conversation is not stored. When you close your browser tab, it is gone. Plain Help has no way to connect any conversation to any person.
What we do log: When you send a message, we record an anonymous event that includes the time, which AI model responded, how long the response took, and whether the response included a reference to a crisis or safety resource. We do not log your message, the AI's response, or any content of the conversation. These logs contain no information that could identify you.
Why we log this at all: The distress tier flag tells our clinical team how often the tool is surfacing crisis resources, so they can review whether those resources remain accurate and appropriate. It is an aggregate signal, not a record of individual conversations.
Third parties: Your message is processed by Anthropic's API to generate a response. Anthropic's privacy policy governs how they handle API data. We use Supabase to store the anonymous event logs described above.
Plain Help will never sell, share, or use data to identify users. There is no advertising, no data brokerage, no research use without explicit community consent.
If you are a domestic violence survivor, an undocumented person, an LGBTQ+ person in a hostile jurisdiction, a human rights defender, or anyone for whom being identified through a tool like this could cause real harm: the architecture above is why this tool was built the way it was. Privacy here is not a feature. It is a safety measure.
Questions: plainhelp.feedback@gmail.com