Privacy policy.
A full privacy policy is in progress. In the meantime, here’s what matters most.
The short version
eviStreams is a research tool built at Penn Dental Medicine. We collect only the data necessary to run the platform — account credentials, uploaded documents, and extraction results. We do not sell your data, share it with advertisers, or use your research content for any purpose beyond what you explicitly request.
Documents you upload are stored securely and accessible only within your project. Extraction results belong to you and can be exported or deleted at any time.
What we collect
Account information
Your email address and hashed password (managed by Supabase Auth). We do not store raw passwords.
Demo visits and optional survey
To start a private demo, we collect your name and email address. You can optionally share your role, interests, and needs. These responses are available only to site administrators, retained for up to 90 days, and can be deleted sooner on request. Follow-up contact requires your separate opt-in. We also record demo uploads and extraction usage against the email you enter, and optional feedback after trying a task. Detailed usage and feedback are retained for up to 90 days. A keyed email identifier and quota counters are retained to enforce the five-extraction and two-upload limits across visits. Demo workspaces have no session timer. After 14 days without activity, the projects in a demo workspace, including any PDFs uploaded to it, are deleted; returning later starts a fresh set of sample projects. The workspace itself is deleted after 90 days without activity. Names and emails are self-reported and are not verified; knowing an email does not grant access to its workspace. Optional product feedback is separate from consent to participate in a formal research study.
Uploaded documents
PDFs you upload for extraction are stored in encrypted object storage. They are used solely to run extraction jobs you initiate.
Extraction data
Form structures, extraction results, and reviewer annotations you create within the platform.
Usage logs
Basic activity logs (job runs, page loads) for debugging and improving the platform. We do not build behavioral profiles.
AI processing
When you run an AI extraction job, document content is sent to third-party large language model APIs to perform extraction. This is limited to the specific job you trigger and governed by the respective provider’s data policies. Do not upload documents containing identifiable patient data without ensuring compliance with applicable regulations (HIPAA, GDPR, etc.).
Questions
For questions about your data or to request deletion of your account and associated data, please contact us via LinkedIn. A comprehensive privacy policy will be published here when the platform moves to broader release.