What Data Does WhatsApp Actually Collect?
What data does WhatsApp collect? Your messages are encrypted, but Meta still gathers metadata, ties you to a phone number, and backups may not be E2E.
By The PrivacyPortal Team
What Data Does WhatsApp Actually Collect?
WhatsApp collects a lot more than most people assume. The content of your messages is end-to-end encrypted by default, so Meta cannot read what you type. But the app still gathers a wide layer of surrounding data — who you talk to, when, from which device and network, and your phone number — and hands much of that context to its parent company, Meta.
That gap between "encrypted content" and "everything else" is the whole story. Let's walk through what data WhatsApp collects, why the encryption headline can be misleading, and what a genuinely private setup looks like instead.
Message content vs. everything around it
It's worth being fair: WhatsApp does use the Signal Protocol to end-to-end encrypt message content by default. Your texts, calls, photos, and voice notes are scrambled on your device and only readable on the recipient's device. Meta does not hold the keys to that content.
The problem is that content is only one part of a conversation. The metadata — the data about your messages — is not protected the same way, and it can be remarkably revealing on its own. Metadata can show:
- Who you message and how often
- When you're active, and for how long
- Which device and app version you use
- Your network and approximate location, via IP address
- Your phone number, and the numbers in your address book
Researchers and intelligence agencies have said for years that metadata alone is enough to map someone's relationships, routines, and social graph — without ever reading a single message.
The data WhatsApp collects, at a glance
Meta's own privacy disclosures describe a broad set of data categories. Here's a plain-language summary of the kinds of information involved.
| Category | Examples | Encrypted end-to-end? |
|---|---|---|
| Message content | Texts, calls, media, voice notes | Yes, by default |
| Account identity | Your phone number, profile name, photo | No |
| Contacts | Numbers you've saved, uploaded to match friends | No |
| Usage & connection | Time online, frequency, device model, OS, signal | No |
| Network data | IP address, mobile network, general location | No |
| Backups | Chat history stored in your cloud provider | Only if you enable encrypted backups |
The key takeaway: the encrypted column has exactly one row. Everything used to identify you, connect you to others, and place you on a network sits outside that protection.
Your phone number is the anchor
WhatsApp requires a phone number to register. That number is not an incidental detail — it's your identity on the platform. It ties your account to a real-world SIM, which usually ties back to a name, a billing address, and a carrier. Even people who never learn your name can often find you by number, and the number links your WhatsApp activity to any other service that knows it.
This is the single biggest difference between "encrypted messaging" and "private messaging." Encryption hides what you said. A required phone number reveals who you are.
Backups are the quiet leak
Here's the detail that surprises people most. Your live chats are end-to-end encrypted, but if you back up your chat history to a cloud provider, that backup is not end-to-end encrypted unless you have specifically turned on encrypted backups. Without it, a copy of your conversations can sit in cloud storage in a form that is far easier to access than the encrypted live chat.
Worse, encryption is a two-sided contract. Even if you disable or encrypt your backups, the person you're talking to might be backing up the same conversation unprotected. Their choice affects your privacy.
Why "encrypted by default" isn't the full answer
None of this makes WhatsApp's content encryption fake — it's real, and it matters. But it means the reassuring phrase "end-to-end encrypted" answers a narrower question than most users think they're asking. It tells you the body of the message is protected. It says nothing about the envelope: the sender, the recipient, the timing, the frequency, and the identity behind the account.
For a deeper walk through why the envelope matters as much as the letter, see our explainer on messaging metadata. It's the piece of the puzzle that determines whether a service is minimizing what it knows about you or quietly accumulating it.
What a lower-data messenger looks like
If the concern is metadata and identity rather than just message content, the fix is to choose tools that are built to know less about you in the first place. A few examples across the space:
- Signal encrypts content by default, is open-source, and works hard to minimize metadata (its "sealed sender" hides who's messaging whom). It still requires a phone number to register, though usernames can hide that number from other users.
- Session drops the phone number entirely and routes traffic through an onion network.
- Threema, a paid Swiss app, also lets you register without a phone number.
- Telegram is a common point of confusion: its cloud chats are not end-to-end encrypted, only its opt-in one-to-one "Secret Chats" are — so it shouldn't be treated as encrypted by default.
Each takes a different swing at the same core problem: reduce the identifying data the service holds.
Where Graphite fits
Graphite is built by PrivacyPortal around the idea that the least private part of most messengers — your identity and network footprint — should be designed away, not just encrypted.
- No phone number, no email. There's no account on the server at all. Your identity is a 12-word recovery phrase generated on your device. There's no number to leak, sync, or link back to you.
- Content encryption with a post-quantum layer. Messages use NaCl/libsodium sealed boxes (X25519) plus a hybrid post-quantum layer (ML-KEM-768), with forward secrecy from a Signal-style double ratchet.
- Tor by default, fail-closed. Graphite bundles Tor and routes signaling and messages through a hidden service, which hides your IP and network — directly targeting the metadata layer that ordinary encryption leaves exposed. Real-time call media runs peer-to-peer over DTLS-SRTP rather than Tor; on mobile, calls are routed so the other person doesn't see your IP.
- A no-logs relay. The relay only ever holds sealed ciphertext, keeps it for a maximum of 7 days, and deletes it after delivery. It never sees your plaintext, your keys, or your recovery phrase.
- On-device features. Message search runs locally — nothing is uploaded. Group chats and file sharing are encrypted too.
Graphite runs on Windows (a portable .exe with bundled Tor), Android (a sideloaded APK that works on GrapheneOS and de-Googled phones with no Google Play Services), and as a web app. It's free and currently in public beta (v1.13.25).
The bottom line
So, what data does WhatsApp collect? Your message content stays encrypted, but Meta still gathers a rich metadata picture — your phone number, contacts, connection details, and usage patterns — and your backups may not be end-to-end encrypted unless you've turned that on. Encryption protects the what; it doesn't hide the who, when, and how often.
If you want a messenger that's designed to hold as little of that as possible, that's the problem Graphite sets out to solve. For a broader comparison of options, our guide to a private WhatsApp alternative lays out the trade-offs.
Try Graphite — free public beta. No phone number, no email, no logs. Available on Windows, Android, and the web. Start at graphite.chat.