Rividor Inc.
Trust Center
Rividor holds safety records, inspection photos, meeting recordings and chemical inventories for frontline teams. This page explains where that data lives, who can reach it, which vendors touch it, and how to reach us when something looks wrong.
Last updated October 19, 2026
Overview
Rividor is an EHS platform for the frontline workforce: forms and inspections, chemical and SDS management, meetings and recordings, and AI that reads what your crews submit. Customers reach it at app.rividor.com (the “Service”). This marketing site, rividor.com, is a separate deployment with no access to customer data.
Everything below describes how the Service is built today. We are an early-stage company; where a control is not yet in place we say so rather than imply otherwise.
Infrastructure & hosting
- Application hosting. The Service runs on Vercel in the United States. Each request is served by short-lived serverless functions; no customer data is stored on application servers.
- Database, auth and files. Customer data lives in a PostgreSQL database, with user authentication and file storage, hosted by Supabase. Recordings, photos and documents are stored in private storage buckets and served through expiring, signed URLs.
- Separate environments. Production and staging run against separate database projects with separate credentials.
- Separate marketing site. rividor.com is deployed on its own, without database credentials or any route into the Service. A compromise of the website cannot reach customer records.
Data protection
- Encryption in transit. All traffic between your browser, the Service and our vendors uses TLS (HTTPS). Plain HTTP is not served.
- Encryption at rest. Our database and file storage providers encrypt stored data at rest.
- Tenant isolation. Every record carries the account it belongs to, and PostgreSQL row-level security policies enforce that boundary inside the database itself, in addition to the application’s own checks. One customer’s query cannot return another customer’s rows.
- Server-side authorization. Writes and sensitive reads go through a server-side service layer that checks the user’s role and permissions before touching data. Browser-side code never holds privileged credentials.
- Rate limiting. Sign-in and the public API endpoints (QR-code form submissions and uploads, SDS requests, tracking links) are rate limited to blunt brute-force and abuse.
- Minimal leakage. Application pages are marked not to be indexed by search engines and send no referrer to third-party sites.
Access & authentication
- Sign-in options. Email one-time code or magic link, email and password (with a verified reset flow), and single sign-on with Microsoft Entra ID for organizations that use it.
- Roles. Within an account, users are members, admins or super-admins. Admins manage people, forms and settings; members see and submit what they are assigned. Permissions are checked on the server for every request.
- Public forms by design. Forms can be published behind a QR code so that workers and contractors submit without an account. These endpoints accept only the fields the form defines and are rate limited.
- Multi-factor authentication. The Service does not yet offer in-app MFA. Organizations that need it should sign in through Microsoft Entra ID, where their own MFA and conditional-access policies apply.
- Rividor staff access. Rividor engineers access production data only to operate the Service or on a customer’s request for support, and production credentials are not checked into source control.
AI & your data
Rividor uses large language models to read uploaded forms and emailed documents, answer questions in the in-product assistant, summarize submissions, and transcribe meeting recordings. This is how that works:
- Only on request. Content is sent to an AI vendor only when a user invokes a feature that needs it: a document is uploaded for extraction, a question is asked, a recording is submitted for transcription.
- Vendors. Text and document understanding uses Anthropic’s Claude models; speech-to-text uses ElevenLabs. Under Anthropic’s commercial terms, inputs and outputs sent through the API are not used to train their models.
- What we log. For each AI call we record the feature, model, token counts, latency and cost in our analytics tool so we can monitor quality and spend. We do not send the content of the call to analytics.
- Human in the loop. Fields extracted from an uploaded or emailed document land as a draft record that a user reviews before it becomes a submission.
Retention & deletion
- Customer records (forms, submissions, chemicals, recordings, transcripts) are kept for as long as the account is active, because EHS records are often required for regulatory look-back periods.
- Transient data is cleaned up automatically on a schedule: temporary uploads that never became a submission, processed inbound email, images removed from documents, and rate-limit counters.
- Deletion requests. Account admins can delete records in the product. To close an account or request deletion of all data, email info@rividor.com. Vendor copies (for example analytics events) are deleted under each vendor’s retention schedule.
Monitoring & reliability
- Error tracking. Unhandled errors on the server and in the browser are captured with full stack traces, so failures surface to the team without customers having to report them.
- Durable background work. Transcription, document parsing and notification delivery run as durable, step-by-step jobs that retry on failure and never silently drop work.
- Field-first by design. Answers and photos in an in-progress public form are cached on the device, so a refresh or a dropped connection does not lose a worker’s draft. Nothing is uploaded until they submit.
Subprocessors
A subprocessor is a third party that processes customer data on Rividor’s behalf. We limit the list to vendors we need to run the Service and this website, and we send each one only the data its job requires. This list is maintained from the code that actually calls these vendors, not from a spreadsheet, and was last reviewed on October 19, 2026.
Service subprocessors (app.rividor.com)
These vendors may receive data you or your team enter into the Service. Rows marked conditional only receive data when a customer uses that feature.
Website subprocessors (rividor.com)
These vendors receive data from visitors to this marketing site: analytics, form submissions and demo bookings. None of them has access to the Service. Vercel, PostHog, Slack (listed above) also serve this website.
The free heat-risk tool at rividor.com/heat sends the location you type to Open-Meteo (geocoding and forecast) and to NOAA’s National Weather Service (active alerts) directly from your browser. No account data is involved.
Changes to this list
We update this page before a new subprocessor receives customer data. To be notified of changes by email, write to info@rividor.com with “Subprocessor updates” in the subject line.
Compliance
Rividor has not yet completed a third-party audit such as SOC 2 or ISO 27001. We answer security questionnaires, sign data processing agreements, and walk through the controls above on a call with your IT or security team. Our Privacy Policy and Terms & Conditions describe our legal commitments.
Report a vulnerability
If you believe you have found a security issue in Rividor, email info@rividor.com with “Security report” in the subject. Include the steps to reproduce and, if you can, the account or URL involved. We will acknowledge your report, keep you informed as we investigate, and will not pursue good-faith researchers who avoid accessing or altering other customers’ data.