Raw activity never leaves the computer. Only derived work knowledge is shared, and only with people allowed to see it. Admin actions land in a tamper-evident audit log, and every person can see who opened their memories. Security and legal teams can answer "who saw what, and when?" from the dashboard.
We're a young company. We hold no certifications and have no audit reports yet. SOC 2 and ISO 27001 audits are planned, not done. Below is exactly where each control stands today.
Each control below says where it stands: Live means it's in the product today; Planned means it is on our roadmap and not available yet. We don't claim what we haven't shipped.
Every question is tied to a real person. Access is checked on every record with eight rules (private, team, department, organization, explicit shares; admins see all). The AI never sees memories the asker can't see.
Raw activity stays on each computer in an encrypted store. Shared knowledge is stored encrypted by Reattend for organizations on reattend.com, and travels over HTTPS. Customer-managed encryption keys are not available; organizations that need full control run their own server.
Every answer names who it came from. Admin actions are written to a hash-chained, append-only audit log: each entry's hash includes the previous entry's, so any tampering breaks the chain.
Only apps and sites marked as work count as work; personal time never leaves the computer. Only derived knowledge is shared, never screen text, audio, transcripts, browsing history, clipboard or documents. The desktop app sends no analytics and does no tracking.
Hosted organizations run on a single server in one United States region, with regular backups kept off-site. We do not offer an uptime SLA today. Each person's raw memory lives on their own computer, not on our server.
Sign-in and API routes are rate limited. We patch known vulnerabilities in our dependencies as we learn of them. There has been no external pen test yet; one is planned.
Here's what happens when work is captured on a computer and a colleague later asks about it. Raw activity never leaves the computer. With your own server, Reattend holds nothing.
Captured on the person's own computer by the desktop app, from work apps only. An AI model turns it into memories: by default Rabbit, our own model, on the computer itself and still in training. With managed AI or the organization's own OpenAI or Anthropic key, the text being written up goes to that AI for the moment it takes; Reattend records usage counts, never the text.
Only derived work knowledge leaves: titles, summaries, facts, tasks, names, meeting notes. Sent over HTTPS and stored encrypted.
Every record is checked against eight access rules before the AI sees it. Only what the asker is allowed to see is used.
Admin actions land in a hash-chained, append-only log, and each person can see who opened their memories. Each entry binds to the previous one's hash, so tampering breaks the chain.
Raw activity stays on each computer, wherever it is. For organizations on reattend.com, shared knowledge is stored encrypted in one United States region. If your law requires a different boundary, or no cloud at all, run your own server: laptops talk only to it, it mirrors the AI models and installers, and Reattend holds nothing. Air-gapped installs are possible.
We have no audit reports yet. The DPA, security overview, subprocessor list and incident response summary are available on request.
Certifications and the pen test are planned, not done. Email anjan@reattend.com if you need any of these for procurement and we'll tell you exactly where each one is.
Straight answers to what security teams ask us. If yours isn't here, write to anjan@reattend.com.
No. We do not train any model on your data. With Rabbit, the desktop app's AI runs on the computer itself. Managed AI and the dashboard's insights use a provider that does not train on your data; with your own key, your agreement with your provider applies. The list of subprocessors is available on request.
Raw activity stays on each computer in an encrypted store and is never uploaded. For organizations on reattend.com, the shared work knowledge is stored encrypted in one United States region, with backups. With your own server, it stays on that server and Reattend holds nothing.
Our target is to notify affected customers within four hours of confirming a P1 security incident, with a written follow‑up inside 72 hours. We don't have a contractual SLA today; one can be agreed in a Government contract. Status incidents (degraded service, not security) post to our status page in real time.
Admins can erase a person's data from the dashboard, and each person can take back any item they shared from the desktop app. The erasure is recorded in the hash‑chained audit log so you can prove it happened. We don't issue a separate cryptographic receipt.
Yes. Government and regulated organizations can run their own server: laptops talk only to it, it mirrors the AI models and installers, and Reattend holds nothing. Air‑gapped installs are possible and scoped with our team. A trainer can be sent to scan existing paper records.
Admin actions are appended to a hash‑chained audit log for your organization. Admins can view it in the dashboard and export it as CSV or JSON, with a chain tip auditors can re-check. Every person can also see who opened their memories. Streaming to an external SIEM is not available; the export is on demand.
Production access is currently held by the founding team only and gated by SSH key + 2FA. Every administrative session is logged. We don't have a formal break‑glass and per‑request approval workflow yet; that's planned as part of SOC 2 readiness. With your own server, we have no access at all.
Access is checked before retrieval: the AI only ever sees records the asker is allowed to read, under eight visibility rules covered by an automated test suite. The AI model in the desktop app runs on the computer. Raw activity is never shared, so it can never reach an answer for someone else.
We notify customers in writing before adding or expanding a sub‑processor. Today that is a direct email to your account contact. The current sub‑processor list is shared on request: email anjan@reattend.com.
None yet. We're a young company; the first external pen test is planned, then at least yearly. Until then we patch known vulnerabilities in our dependencies and rate limit sign-in and API routes.
Email anjan@reattend.com with your assessor and what you need to see. We don't have a packaged trust report or pre‑mapped CAIQ / SIG Lite responses yet. At our stage, the best path is a 30‑minute call where the founder walks you through the architecture, controls, and current audit status. We reply within 12 hours, every day.
Compliance is a conversation, not a checklist. If your security or legal team needs something we haven't published, such as a control mapping, a draft DPA or a deeper architecture review, we'd rather talk to you directly than write it on a marketing page.
