Trust & Security
Your relationships are sacred. So is your data.
Authentic CRM handles dozens of real relationships, not millions of leads. Fewer records, more care. Here is exactly how we protect your data.
Infrastructure
Encrypted in transit
All connections use TLS encryption. We enforce HTTPS-only connections so your browser never falls back to an unencrypted channel.
Encrypted at rest
Your data is stored on Neon PostgreSQL, which encrypts all data at rest. Neon maintains SOC 2 Type II compliance. Credentials for the accounts you connect (Gmail, Slack, Granola, Mailchimp) and two-factor secrets are encrypted a second time by the application, with a key held outside the database.
Hosted on Vercel
Our application runs on Vercel's edge network with built-in DDoS protection. Vercel maintains SOC 2 Type II compliance. See security.vercel.com for details.
Protected at the edge
Vercel's firewall rate-limits sign-in and our public endpoints before a request reaches the application, and logs automated traffic for review.
Security headers
We set strict Content Security Policy, X-Frame-Options, and other headers to protect against clickjacking, XSS, and content injection.
Data isolation
Organization-level separation
Every database query is filtered by your organization. Your data is never accessible to other accounts. Automated checks run on every code change to verify that each server action authenticates its caller and each owner-only action checks the caller's role, and we audit the full surface by hand each quarter.
Role-based access
Organization owners control who has access and who can export data, delete the workspace, or manage integrations. Team members can only see data within their own organization.
Private touches
Owners choose, per source, whether synced email, texts, and meeting transcripts are visible to the whole team or only to the person who captured them. Private contents stay out of teammates' views, shared AI memory, and other people's exports. That a touch happened stays visible, so cadences and dashboards keep working.
Where you're signed in
Sessions expire after 30 days of inactivity and are invalidated on logout. You can see every active session, with its device and approximate location, on your profile and sign out of all of them at once. Sign-ins from a new country are flagged in your workspace's security log.
Rate limiting
Sign-in, API, and public endpoints are rate-limited at the edge and again in the application to prevent brute-force and scraping attacks.
AI and your data
Your data never trains AI models
We use Anthropic's Claude for AI features. Anthropic does not use API inputs to train their models. Your notes and context are processed, not learned from.
You control what AI sees
AI suggestions are generated from the notes and context you've entered. You decide what to capture, and AI only works with what you've provided.
AI suggests, never acts
AI proposes touch suggestions with reasoning you can review. It never sends messages, creates records, or takes action on your behalf.
Transparent reasoning
Every AI suggestion includes an explanation of why it was made, so you can evaluate the recommendation before acting on it.
You approve connected assistants
When you connect Claude to your CRM, it can read your relationships and pipeline to answer questions. Only if you separately grant write access can it create people, touches, tasks, and opportunities — it can never send, delete, or edit anything, every write is listed in Settings, and each one can be undone. Connecting requires you to sign in and explicitly grant access, and you choose which workspace.
Connections are yours, and revocable
Each connection is tied to your own account, is never shared between people, and can be disconnected instantly from Settings.
Authentication
Password security
Passwords are securely hashed before storage using industry-standard algorithms. We never store or log plaintext passwords.
Two-factor authentication
Add an authenticator app to your account, with recovery codes as a backup. Workspace owners can require a second factor for every member of the team.
Passkeys
Sign in with Touch ID, Face ID, Windows Hello, or a hardware security key instead of a password. Passkeys are phishing-resistant: they only work on our domain, so a look-alike site cannot use them.
Bot protection
Login and signup are protected by Cloudflare Turnstile to prevent automated attacks.
No third-party tracking cookies
We use a single, essential session cookie for authentication. No advertising or third-party tracking cookies.
API key security
API keys are scoped to what each integration needs, securely hashed before storage, and can be revoked at any time from Settings.
Monitoring and response
Continuous monitoring
Failed sign-ins, rate-limit trips, and blocked outbound requests raise an alert within minutes. A weekly security digest is reviewed by a person, not just filed.
Audit trail
Owners can see a log of security and admin events in Settings: sign-ins from new locations, two-factor and API key changes, team changes, and every write a connected assistant makes. Security events are kept for 90 days and admin events for a year.
Incident response
We keep a written incident response procedure covering containment, credential rotation, and notification. If an incident affects your data, we tell you what happened and what we did.
Availability
The application is monitored externally around the clock, and errors are reported to us with personal data stripped.
Payment security
Stripe handles payments
All subscription payments go through Stripe, which is PCI DSS Level 1 certified. We never see or store your full card number.
Gift purchases
When you buy a gift through Goody, card details are entered in a Very Good Security (VGS) hosted form and tokenized before they reach Goody. They never pass through our systems.
Companion apps
Mac text sync
Our Mac app syncs the text of your iMessage conversations — but only with people already in your CRM, matched on your device, so nothing else ever leaves your Mac. Calls sync as who-and-when only. Synced texts are stored like your emails (encrypted at rest), follow your workspace's private-by-default setting, and can be kept private to you.
What the extension accesses
The Chrome extension reads publicly visible information from LinkedIn profile pages you visit — name, headline, company, location, photo, and about section. It only reads data when you click "Add to Authentic."
Data stays in your CRM
Profile information is sent directly to your Authentic CRM account over an encrypted connection. It is not shared with any other service or third party.
No background data collection
The extension does not monitor your browsing activity, collect data from other websites, or run in the background. It only activates on LinkedIn profile pages.
Local storage
The extension stores only your API key locally in Chrome's sync storage to authenticate with your CRM account. You can revoke the key at any time from Settings.
Your rights
Your data is yours
You own the data you put into Authentic CRM. We do not sell your data to third parties, and we never will.
Data export
Organization owners can export all data at any time from Settings. The export is a JSON file containing your people, companies, touches, tasks, milestones, opportunities, knowledge items, and more.
Account deletion
Organization owners can delete their organization and all associated data at any time from Settings. Deletion is immediate and permanent.
GDPR
We respect your data protection rights including access, correction, deletion, and objection. Contact us to exercise any of these rights.
Sub-processors
These are the third-party services that process data on our behalf. Each is selected for its security posture and data handling practices. Services marked “only when you connect it” receive nothing unless you turn them on.
| Service | Purpose | Data processed | Notes |
|---|---|---|---|
| Anthropic | AI suggestions, memory, and connected Claude assistants | Contact notes, message content, relationship context | Does not train on API data |
| TypeSafe | AI classification (reply detection, relationship type) | Message and relationship context, returned as scores only | — |
| Vercel | Hosting and edge firewall | Application data, request logs | SOC 2 Type II |
| Neon | Database | All application data, encrypted at rest | SOC 2 Type II |
| Upstash | Rate limiting and alert state | Hashed identifiers and counters only | — |
| Sentry | Error monitoring | Error reports with personal data stripped | — |
| UploadThing | File storage | Files you attach to touches and documents | — |
| Cloudflare Turnstile | Bot protection on sign-in and signup | IP address and browser signals | — |
| Resend | System emails (invites, resets, digests) | Email address, message content | — |
| Stripe | Subscription payments | Billing name, tokenized card data | PCI DSS Level 1 |
| Goody | Gift sending | Recipient name and shipping address, gift details | Only when you send a gift |
| Gmail and Google Calendar | Email content, calendar events | Only when you connect it · SOC 2 Type II | |
| Nylas | Microsoft 365 and other mailboxes | Email content, calendar events | Only when you connect it |
| Granola | Meeting transcripts | Meeting notes and transcripts | Only when you connect it |
| Slack | Digests and notifications | Summaries you choose to send to your channels | Only when you connect it |
| Mailchimp | Audience sync | Names and emails of the contacts you choose to sync | Only when you connect it |
| Perplexity | News detection | Contact and company names | Only when you turn it on in Settings |
| Exa | Company and industry research | Company names and websites | — |
| Apollo, People Data Labs, EnrichLayer | Contact enrichment | Name, email, LinkedIn URL | Per contact, on request |
| Apify, BrightData | Public LinkedIn posts | Public LinkedIn profile URLs | — |
| RB2B, IPinfo | Website visitor identification | Visitor IP address and company | Only if you install the pixel |
| Mapbox | Address autocomplete | Address text as you type it | — |
Questions about security?
We are happy to discuss our security practices, a Data Processing Agreement (DPA), or anything else your review needs. If you believe you have found a vulnerability, please report it to the same address; we read every report and will tell you what we did about it.