PII redaction, by default
Emails, phone numbers, national IDs, and card numbers typed into chat are masked before they’re stored. On every tier.
- [EMAIL]
- [PHONE]
- [NATIONAL_ID]
- [CARD]
Product Security & trust
Exactly how OptoBot stores, protects, and deletes data, stated plainly, including what we haven’t built yet.
What the visitor types
Hi, I’m Amira. Call me on 0100 555 0142 or email [email protected]
What OptoBot stores
Hi, I’m Amira. Call me on [phone] or email [email]
On by default, on every tier. Sample values, nothing is sent.
OptoBot keeps your content, conversations, and leads in its own database and searches your content in place, with PII redaction applied before anything is stored.
OptoBot database
Searched in place
Emails, phone numbers, national IDs, and card numbers typed into chat are masked before they’re stored. On every tier.
Paying organizations’ data is never purged; dormant free-tier projects are anonymized after 5 years of inactivity.
A record of administrative actions, included from the BUSINESS plan.
Opt-in authenticator-app 2FA with recovery codes and lockout, or sign in with Google.
A token minted for one job can’t be replayed for another.
PII redaction is a platform default on every tier. The audit log is surfaced from BUSINESS.
Trust runs in both directions: into the knowledge base, and out of every answer.
We won’t dress the product up with badges it doesn’t have. Here’s the ledger.
The privacy policy spells out what we process and why. Read it
On the roadmap, not available today
Ask us where they stand. The answer changes, and this page will change with it. Talk to us about Enterprise
Plain answers about how OptoBot works. If yours isn’t here, ask the people who build it.
In OptoBot's own database. Your documents, search index, conversations, and leads stay in that one store, and PII is redacted before anything is stored. No external search service holds a copy.
Retention is a built-in policy, not a setting you have to negotiate: active projects and paying organizations' data is never purged, and dormant free-tier projects are anonymized after 5 years of inactivity. Personal information is redacted from stored transcripts either way.
Inbound Meta events are signature-verified before processing, your connection credentials are AES-256-GCM encrypted at rest, and duplicate webhook redeliveries can't answer twice.
Only on origins you allowlist. The widget uses short-lived embed tokens scoped to your project, its frame-ancestors CSP restricts embedding, and you can rotate the embed key at any time.
No — and we won't claim any. What's built in: PII redaction, automatic retention, audit logging (from BUSINESS), 2FA, and hardened embeds. This page and the privacy policy describe exactly that.
On the roadmap, not shipped. See "On the roadmap" above — or book a demo and ask us directly.
We showed the mechanisms. Start free, or bring this page to a demo and quiz us.