Sovereign Cloud now in early access

Make every document you own searchable with AI. Without giving it away.

Connect your OneDrive, Dropbox, or Google Drive. We import, embed, and make every file searchable using our patent-pending ZH-embeddings engine. Zero hallucinations. Encrypted with your key. Your data never trains anyone's model.

Import OneDrive, Dropbox, Google Drive Patent-pending ZH-embeddings Zero hallucinations

OneDrive holds your files. ChatGPT reads them. Neither makes them yours.

Cloud storage was built to hold dumb files. It does not make them searchable in any meaningful way. So you copy them into ChatGPT or Claude or Gemini to find what you need. And in that moment, your contracts, research, patient files, and inventions become part of someone else's data infrastructure.

For a healthcare org with patient records and research IP. A law firm with privileged communications. A family office with client documents. A founder with strategic plans. A defense contractor with regulated data. That tradeoff is unacceptable.

Search My Data was built to replace both: the dumb cloud storage AND the hallucinating chatbot.

The difference

  • OneDrive, Dropbox, Google Drive: dumb file storage you rent
  • ChatGPT, Claude, Gemini: your data on their servers
  • Generic AI search: hallucinations and made-up citations
  • Search My Data: import everything you have, make it searchable
  • Search My Data: encryption with your key, not ours
  • Search My Data: deterministic answers verified against your documents

How it works

Upload, search, and verify. All without your data leaving your control.

1

Upload

PDFs, Word docs, images with OCR, spreadsheets. Encrypted before they leave your device.

2

Index

Local semantic embedding with 4096-dimensional vectors. No third-party model APIs.

3

Search

Ask questions in plain English. Get answers verified against your own documents.

4

Verify

Every answer cites the exact source document and page. No hallucinations. Ever.

PATENT-PENDING ZH-EMBEDDINGS

Bring your entire OneDrive, Dropbox, or Google Drive. Make all of it searchable.

Cloud storage was built to hold files. We make those files intelligent. Connect your account, we import everything, and our patent-pending ZH-embeddings engine makes your data searchable with deterministic AI.

  • One-click import. OAuth into your OneDrive, Dropbox, Google Drive, or SharePoint. We handle the rest.
  • Folder structure preserved. Your hierarchy comes with you. Permissions tracked. Audit log starts immediately.
  • Patent-pending ZH-embeddings. Our proprietary embedding pipeline produces 4096-dimensional vectors that make YOUR data searchable with AI. No third-party model APIs.
  • All file types. PDFs, Word, Excel, PowerPoint, images with OCR, emails, archives. We handle the formats your cloud storage struggles with.
  • Incremental sync (coming). Keep using OneDrive while we mirror it. Or migrate fully when you are ready to leave.
Import status
📁
OneDrive
612 GB · 47,832 files
✓ Done
📁
Dropbox Business
183 GB · 12,447 files
✓ Done
📁
Google Drive
94 GB · 8,221 files
Embedding...
ZH-embeddings progress 68,520 / 68,500 files
~12 minutes remaining

Three ways to deploy

Pick your level of sovereignty. Start in the cloud, move to hardware when you want it in your hand.

SHIPS IN 4-6 WEEKS

Sovereign Drive

Encrypted hardware drive

An encrypted SSD pre-loaded with the full search engine. Plug into any Mac or Windows machine. Works completely offline.

From $1,500 + $99/mo software
  • 2TB to 8TB capacity options
  • Mac and Windows compatible
  • Hardware-encrypted enclosure
  • Local UI, no internet required
  • Data never leaves the drive
  • Buy 2 for redundancy ($500 off second)
Reserve now
SHIPS IN 4-12 WEEKS

Sovereign Bundle

Drive + dedicated host server

Sovereign Drive plus a pre-configured Mac server, customized to your needs. Plug in power and ethernet. Built for teams from 5 to 50+ users.

From $3,500 + $99/mo software
  • Pre-configured Mac server
  • Sized for your team and document volume
  • Multi-user access via local network or Tailscale
  • Redundant server option available
  • White-glove setup included
  • Full audit log and admin panel
Talk to sales

Why Search My Data

Built by engineers who deploy sovereign systems for nation-states. Now available for the rest of us.

🎯

Zero hallucinations

Powered by the ZH Standard deterministic verification engine. Every answer is verified against your documents. No made-up citations. No fabricated facts. Every claim points to a real source.

🔐

Your data stays yours

No OpenAI, Anthropic, or Google APIs. We use open-source, pre-trained language models combined with ZH Standard verification, all running on infrastructure we control. Your documents never train anyone else's AI.

🔑

You hold the key

Customer-side encryption. We hold your encrypted bytes. We literally cannot decrypt them. If you lose your key, we cannot reset it for you. A reset would imply we can access your data. We cannot. Real sovereignty has consequences. We embrace them.

🏛️

Sovereign by design

Sovereign Cloud is dedicated to you alone. Sovereign Drive runs entirely on your hardware. You decide where your data lives and who can touch it.

📋

Audit everything

Every search, every result, every document accessed is logged. You control the audit trail. Required for HIPAA, SOC 2, and ethical wall compliance.

⚙️

Built by operators

Brightstead Technologies builds deterministic verification systems for sovereign nations and Fortune 500 companies. Search My Data is the personal-scale version of what we build for governments.

Built for organizations whose documents matter

If your work is too valuable or too regulated to feed into a chatbot, this is for you.

Healthcare and biotech

Patient records, clinical trial data, FDA submissions, research IP. HIPAA-compliant audit trails. Search across years of work without exposing protected health information to model training.

"I have inventions and research I will never put into ChatGPT. I needed something I could trust with my actual work."

Law firms and legal teams

Privileged communications, case files, discovery materials. Maintain conflict-of-interest walls. Audit access for ethics compliance. Compartmented permissions support ethical wall requirements.

"Westlaw and Lexis cost more and the data still lives somewhere else. This is half the price and stays with us."

Financial services and family offices

Investor documents, due diligence, deal flow, LP communications. SEC and FINRA compliance. Confidential client records. No data shared with third-party model providers.

"Our clients expect their information to stay private. This is the only AI tool I can use without breaking that."

Founders and executives

Board materials, strategic plans, M&A documents, IP. Search your knowledge base without giving competitors a window into your strategy. Cap tables, term sheets, and trade secrets stay yours.

"I am not putting our cap table or term sheets into a public AI tool. Period."

Government and defense contractors

ITAR, CUI, FOUO, and classified-adjacent work. On-premises deployment available. Air-gapped operation supported. Custom compliance requirements welcome.

"We needed something that could run completely offline. The Sovereign Drive is exactly that."

Energy, pharma, and research orgs

Seismic data, clinical trials, grant-funded research, patent prosecution. Search across decades of institutional knowledge. Protect IP that took years and millions of dollars to develop.

"I have 20 years of research across hundreds of projects. Now I can actually find things without giving them away."

The trust architecture

Sovereignty is not a marketing word. It is something we built into the system you can verify.

Customer-side encryption

Your encryption key is generated on your device. We never see it. Not in transit. Not in storage. Not in logs. We hold encrypted bytes that look like static noise.

Verifiable telemetry

For Sovereign Drive, our health monitoring sends only system metrics: CPU, disk, uptime. No document data. No queries. The telemetry code is open for inspection.

Open-source models, no third-party APIs

We use open-source, pre-trained language models combined with the ZH Standard verification engine. No calls to OpenAI, Anthropic, Google, or any frontier model API. All reasoning runs on infrastructure we control. Your queries and documents never leave the system.

Where compute happens

Persistent data — your documents, embeddings, audit log, account — lives only on infrastructure we own. Inference compute runs on our hardware by default. When demand exceeds local capacity, we route to dedicated GPU pods that are network-locked, ephemeral, and see only the chunks for one query. Sovereign-Only mode (premium) keeps everything on owned hardware.

Tenant isolation

Sovereign Cloud customers get dedicated database schemas with row-level security. Your data is never colocated with another customer's data in shared tables.

Audit-grade logging

Every search, every document view, every export is logged with cryptographic chain of custody. SHA-256 audit chains. You control the audit trail.

The ultimate proof

If you lose your encryption key, we cannot reset it for you. A reset would imply we can access your data. We cannot. Your data becomes permanently inaccessible, including to us. That is the proof we cannot decrypt it.

Simple pricing for every team size

Start with Sovereign Cloud today. Upgrade to hardware when you want maximum sovereignty.

Solo

$99/month
1 user · 50 GB storage
  • Dedicated encrypted instance
  • 10,000 searches per month
  • All file formats supported
  • Email support
  • 14-day free trial
Start free trial

Practice

$1,200/month
Up to 20 users · 1 TB storage
  • Everything in Team
  • Unlimited searches
  • Compartmented permissions (coming soon)
  • Audit log dashboard
  • Phone and email support
Start free trial

Enterprise

Custom
Unlimited users · Custom storage
  • Everything in Practice
  • SSO and identity integration
  • Custom deployment options
  • SLA and dedicated support
  • White-glove onboarding
Contact sales

Annual billing: 2 months free (pay 10x the monthly rate). Add additional users at $50/month each.

2TB Drive

$1,500 + $99/mo
~500K pages of documents
  • Hardware-encrypted enclosure
  • Mac and Windows compatible
  • Local UI, works offline
  • Software updates included
  • Solo or single-user use
Reserve now

6TB Drive

$2,500 + $99/mo
~1.5M pages of documents
  • Everything in 4TB
  • For large document sets
  • Practice tier ready
  • Multi-user team capable
  • Up to 5 concurrent users
Reserve now

8TB Drive

$3,000 + $99/mo
~2M pages of documents
  • Everything in 6TB
  • Maximum capacity
  • Built for firms and groups
  • Up to 20 concurrent users
  • Best for Practice tier
Reserve now

All drives include hardware-encrypted enclosure, Thunderbolt 4 / USB-C, and pre-loaded Search My Data software. Buy a second drive for redundancy: $500 off.

Team Bundle

$3,500 + $99/mo
Up to 5 users · Single server
  • 2-4TB Sovereign Drive
  • Pre-configured Mac server
  • Sized for small teams
  • Local network access
  • White-glove setup
  • Ships in 4-6 weeks
Talk to sales

Firm Bundle

$54K+ + $2,500/mo
Up to 50 users · Mac Studio cluster
  • Multi-drive RAID configuration
  • 2x Mac Studio M3 Ultra (256GB RAM)
  • Apple Business procurement
  • Multi-server cluster
  • SLA and dedicated support
  • Ships in 1-3 months
Talk to sales

Enterprise Bundle

From $120K + monthly
50+ users · Custom architecture
  • 2-3x Mac Studio M5 Ultra (when available)
  • Custom storage configuration
  • On-site setup available
  • Custom integrations and SSO
  • Dedicated account team
  • Lead time: 1-3 months, subject to inventory
Contact sales

All bundles include drive(s), pre-configured Mac server hardware, shipping, and white-glove setup. Hardware lead times reflect current global Apple Silicon inventory constraints. Encrypted Vault remote backup available as add-on from $100/month.

For organizations evaluating the alternatives

A 50-person team running this much compute on cloud APIs (OpenAI, Anthropic) typically spends $1,700-8,500/month, with their data flowing to a third party. Self-hosting the same workload on cloud GPUs (AWS H100 instances) costs $5,800-11,500/month, plus the engineering team to manage it.

The Firm Bundle is owned hardware, sovereign data, and zero ongoing cloud compute exposure. Most organizations see payback in 12-18 months versus their current alternatives. Use the calculator below to estimate your specific savings.

See your needs, costs, and savings

Tell us about your team. We'll recommend a tier and show what you'd save versus cloud alternatives.

1 50 100 250 500+
500 GB 2 TB 5 TB 10 TB 20 TB+
Light Standard Heavy Enterprise

~10 queries per user per day. Suitable for most office work.

Adds a second host server and a second drive for failover. Recommended for any team that depends on uninterrupted access.

RECOMMENDED TIER
Sovereign Cloud Team
A dedicated encrypted instance for up to 5 users with 250GB storage. Best for small teams getting started.
Hardware (one-time) $0
Software / hosting (monthly) $399/mo
Year 1 total $4,788
VS CLOUD ALTERNATIVES
Comparable cloud LLM API spend $2,500/mo
Estimated savings (Year 1) $25,000+
Talk to sales about your build

Estimates based on typical workload patterns. Custom builds for organizations beyond 100 users may include sovereign data center partnerships, custom server architectures, or distributed deployments. Actual quote may vary based on your specific compliance, integration, and infrastructure requirements.

Common questions

Everything you need to know before you start.

Where does the AI compute actually happen?

Persistent customer data — your documents, embeddings, audit log, and account information — lives exclusively on infrastructure we own and operate. It never leaves. Inference compute (the math that turns your question into an answer) runs on our hardware by default. When demand exceeds local capacity, we route to dedicated GPU pods provisioned through our compute partner (RunPod). These pods are network-locked to our infrastructure, use ephemeral storage only, and explicitly clear memory between requests. The compute partner sees only the specific text chunks needed for one query, never your full library, identity, or history. Customers who require zero third-party compute can opt into Sovereign-Only mode as a premium tier.

What is Sovereign-Only mode?

A premium tier where 100% of your inference compute stays on infrastructure we own. We never route any of your queries to third-party GPU partners, even under heavy load. During peak demand, your queries may be slightly slower as they wait for local capacity. Best for healthcare orgs, defense contractors, regulated finance, and anyone whose compliance posture requires zero external compute. Available on Practice tier and above, or as a $200-500/mo upcharge on Solo and Team plans.

Yes. Connect your account via OAuth and we will import your entire file structure (or any subset you choose). Folder hierarchy is preserved, ZH-embeddings are generated for every file, and the original files are stored encrypted on your dedicated instance. Imports can be one-time or ongoing sync (sync is rolling out in early access).

Do I have to leave OneDrive or Dropbox to use Search My Data?

No. You can keep using your current cloud storage and use Search My Data as the intelligent layer on top. Many customers start with a one-time import to evaluate the system, then decide whether to migrate fully. Some keep both indefinitely. The choice is yours.

What does "patent-pending ZH-embeddings" mean?

Our proprietary embedding pipeline uses a 4096-dimensional vector representation combined with the ZH Standard verification framework. The patent-pending technology produces searchable representations of your data that work specifically with our deterministic reasoning engine. Your data is embedded on our infrastructure (Sovereign Cloud) or your hardware (Sovereign Drive). No third-party model APIs are used in the embedding process.

How is this different from ChatGPT or Claude?

ChatGPT and Claude are general-purpose AI tools that send your data to their servers and may use it to improve their models. Search My Data is a dedicated document search system. Your data stays on a dedicated instance or your own hardware. We never use frontier model APIs. Answers are deterministic and verified against your documents only.

How can you guarantee zero hallucinations?

Our search uses the ZH Standard deterministic verification engine. Every answer is generated by retrieving relevant chunks from your documents, then verifying that the response is supported by those exact passages. If a claim cannot be verified, the system will not generate it. Every answer cites the specific source document and page number.

What AI models power Search My Data?

We use open-source, pre-trained language models combined with the ZH Standard deterministic verification engine, plus our patent-pending ZH-embeddings module for vector search. The models run on infrastructure we control. No calls to OpenAI, Anthropic, Google, or any commercial frontier model API. Your data never goes to a third-party AI provider.

What does "encrypted with my key" actually mean?

Your encryption key is derived from a passphrase you set during onboarding. The key never leaves your device. We store your data as encrypted bytes that we cannot read without your key. If you lose your key, we cannot reset it for you. A reset would imply we can access your data. We cannot. Save your recovery phrase somewhere safe (a password manager, a safe, a bank deposit box). If you lose both your key and your recovery phrase, your data becomes permanently inaccessible, including to us. That is the proof we cannot decrypt it.

What is the difference between Sovereign Cloud and Sovereign Drive?

Sovereign Cloud is a dedicated encrypted instance running on Brightstead infrastructure. We hold encrypted bytes, you hold the key. Sovereign Drive is a hardware-encrypted SSD that ships to you and runs on your own Mac or PC. With the Drive, your data physically never leaves your office. Cloud is faster to start; Drive is the maximum sovereignty option.

Can my whole team use it?

Yes. Sovereign Cloud Team plan supports up to 5 users for $399 per month. The Practice plan supports up to 20 users for $1,200 per month. For Sovereign Drive, the Bundle includes a Mac mini host machine that team members connect to over the local network or via Tailscale mesh. Compartmented permissions (ethical walls for legal use) coming soon.

What file formats do you support?

PDFs (including scanned PDFs with OCR), Microsoft Word documents, plain text files, Markdown, Excel and CSV, images with OCR (PNG, JPG, TIFF), and email exports. We are adding more formats based on customer requests.

What happens if I cancel?

You can export all your data and your audit log at any time. After cancellation, your encrypted instance is retained for 30 days in case you change your mind, then permanently destroyed. For Sovereign Drive, you keep the hardware. The software license can lapse, but your data stays accessible (with limited features).

How long until my Sovereign Drive or Bundle ships?

Sovereign Drives ship in 4 to 6 weeks from order. Sovereign Bundles ship in 4 weeks (Team) to 3 months (Firm and Enterprise tiers), depending on your configuration and current global Apple Silicon inventory. The current memory shortage is making procurement slower than usual; reserve early to lock in your delivery slot. Sovereign Cloud is available immediately while you wait.

Start with Sovereign Cloud today.

Upgrade to Sovereign Drive when it ships. 14-day free trial. No credit card required to start.

Questions? Email hello@searchmydata.com or reply to any of our outreach.