Why we built Blockd Home
Trust is the unsolved layer of personal privacy. Encryption was step one. Hardware ownership is the endpoint. Why we built Blockd Home — and what it means for the next surveillance frontier.

I started Blockd because every privacy product I evaluated still required me to trust someone.
That sentence sounds simple. The implications are not.
Signal protects message content with encryption that is mathematically excellent and provably correct — and asks you to register with a phone number that ties every conversation back to a state-issued identity. WhatsApp uses the same encryption protocol Signal pioneered and routes every message through Meta's servers. iMessage encrypts your messages and then quietly decrypts them at Apple's discretion via iCloud Backup. Proton Mail is end-to-end encrypted and lives entirely on Proton's hardware in Switzerland. Telegram's “secret chats” are encrypted but the default mode is not, and the company has operated for a decade with a “we'll figure it out later” posture toward government cooperation.
Every one of these products is good at what it does. None of them solve the actual problem. The actual problem is that trust is the unsolved layer of personal privacy, and encryption — as remarkable an engineering achievement as it is — was only ever step one.
Think about what your phone actually tells the world right now:
- Your phone number, which is tied to your real-world identity in every government database
- Who you message, when, from where, how often
- The IP addresses of the relays and servers your messages traverse
- The cloud accounts that hold backups of those messages
- The metadata embedded in every photo you send
- The voice samples in every call you make
- The AI prompts you type into every assistant you use
None of that is “encrypted content.” It is everything else. And it is the part that actually identifies you, locates you, profiles you, and ultimately sells you to whichever advertiser, employer, government, or insurance company is paying for the data.
You can build encryption on top of broken trust assumptions and still leak everything that matters.
The trust minimization frame
When my co-founder Fazri and I started Blockd in October 2024, we framed the problem differently. Instead of asking “how do we encrypt this conversation better,” we asked: what would it take to require trust from as few parties as possible?
That reframe changes the entire architecture.
Mainstream messengers ask you to trust your phone carrier (because you registered with your number), a software company (because they hold your account), their infrastructure (because they own the servers), their employees (because someone has access), their lawyers (because they comply with subpoenas), their backup providers (because they hold copies), and the operating system underneath everything else. That is seven layers of trust before a single message gets sent.
Blockd strips those layers one at a time.
You do not trust your phone carrier, because Blockd does not require a phone number. You do not trust a backup provider, because Blockd does not have one. You do not trust the company with your social graph, because we use blinded routing tokens — even with full visibility into our infrastructure, an attacker cannot reconstruct your network. You do not trust the company with your message content, because end-to-end encryption was step one, and we use it properly. You do not trust your IP address being seen by our relays, because the Blockd Premium tier routes through Tor before reaching them.
That is the application layer. Most privacy products stop here and call it done. We did not, because there is one layer left.
You still have to trust us — Block Enterprises, the company — to operate honest relays. Our relays cannot read your messages. They cannot see your social graph. They cannot decrypt your content. But they exist. They are operated by us. And no matter how careful we are about logs and metadata, the existence of our infrastructure is a trust requirement.
The only way to remove that last layer is to put the relay in your house.
Why hardware
Blockd Home is the answer to the trust problem at the infrastructure layer. The relay your messages flow through is a device you own, running in your house, behind your network, under your physical control. We can be served a subpoena tomorrow and the response is opaque ciphertext addressed to routing tokens we cannot link back to you — because we never had your messages in the first place. They went through your relay, not ours.
The implications run deeper than messaging.
If you have a device in your house that runs your privacy relay, that same device can run your encrypted file storage, your password manager, your private AI inference, your encrypted email server, your media sync. Every one of those categories is currently dominated by a product that asks you to trust a cloud company. None of those companies are run by villains. They are simply structurally incapable of giving you what you actually want, which is for no one to be in the path.
The hardware version of trustless privacy is not a single device. It is a foundation. The architecture that makes messaging trustless is the same architecture that makes everything else trustless. Blockd Home V1 launches with the messaging relay because that is the use case with the most existing trust pain. The same hardware runs Drive, Vault, Intelligence, and Mail as software updates over the following eighteen months — free to every backer who buys V1.
Private AI is the next privacy frontier
There is one category in this ecosystem where I think the gap between what users want and what the market provides is bigger than anywhere else: private AI.
Every consumer AI product available today sends your prompts to a remote data center. Every conversation with ChatGPT, Claude, Gemini, Copilot — every document you summarize, every email you draft, every question you ask — leaves your device and lands on someone else's hardware. The companies promise they do not use the prompts for training. Some do anyway. The companies promise their data retention limits are short. Some are vague. The companies promise they would never hand the data over to a government. Some already have.
We are about to live through an inflection where AI assistants become as central to daily life as messaging. And we are watching the privacy infrastructure to support that future not get built. The same surveillance economics that made every messaging app a metadata farm are about to make every AI assistant a thought farm — and the consumer market does not yet have an answer.
Blockd Home runs a dedicated Hailo-8 neural processing unit at 26 TOPS. That is enough to run modern open-source language models locally at conversational speed. The thing you type into the assistant on your phone reaches the assistant on your device, in your house, and never leaves the network you control.
That is the version of AI we want to live with. It is the version we are going to build.
What Blockd Home is not
I want to be honest about the limits, because the privacy market is full of products that are not.
Blockd Home is not magic. If your phone is compromised at the operating-system level, no relay architecture in your house can save you. If you take a screenshot of a sensitive conversation and send it to someone you should not trust, no encryption protects you from the human in the loop. If a nation-state-level adversary is dedicating real resources to surveilling you specifically, you need an operational security strategy that goes far beyond hardware ownership.
What Blockd Home does give you is the strongest reduction in metadata exposure that consumer hardware can deliver today. It eliminates the dependency on phone-number identity, on cloud-hosted relays, on backup providers, and on the entire economic model that makes mainstream messengers leak metadata in the first place.
It also gives you something more important than any specific privacy guarantee. It gives you a foundation you own. Every future privacy feature in the Blockd ecosystem runs on the device you bought on day one. No new subscription. No new hardware. No new surveillance.
Why now
We have been at this for close to two years. The Blockd messaging app entered public TestFlight beta on May 1, 2026. A seven-person team — me, my co-founder Fazri Zubair, and a five-person engineering team from our strategic partner Titan Games — has been building the platform full-time since June 2025. We are not a six-month MVP being rushed to market on a hype cycle. We are eighteen months into building the privacy stack we wanted to use ourselves.
Blockd Home reaches Crowd Supply in Q4 2026. If you have read this far and what we are describing resonates — that the trust layer is the unsolved problem, that hardware ownership is the natural endpoint, that private AI is the next surveillance frontier — you are the audience we built this for.
Reserve your spot for Blockd Home when the campaign opens. Or download the Blockd app on TestFlight today and experience the architecture we are betting the company on.
We built Blockd Home because no one else was building it. The reason we built it is the same reason you would buy one.
The privacy you have is not the privacy you think you have. The architecture that fixes it is the one you own.
— Josh Hartsel, Founder & CEO, Block Enterprises (Blockd)

.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)