Privacy Policy

Who we are and how to reach us

UnblockedOps LLC ("UnblockedOps", "we", "us") is a New York limited liability company, and today it is one person. The work is reading security reviews, questionnaires, and architecture for teams trying to get an AI system through security sign-off, which means people send us documents that matter to them. This policy covers this website, booking a First Look, and the material handled under an engagement.

We decide what is collected, where it is held, and when it is deleted. The companies listed under "Who processes it" below handle what they touch on our instructions, under their own data-processing terms.

Privacy questions and requests go to [email protected]. There is no account to log into and no form to fill in.

Information we collect

Website visits

These are static pages. Cloudflare serves the site and keeps the ordinary server and security logs every host keeps: the address a request came from, the browser string, the page asked for, the time it was asked for. Those exist so the site stays up and unabused.

We also measure how the site gets read, with two tools: PostHog and Cloudflare Web Analytics. Whether either of them runs is decided by country. If the connection is coming from the EU, the EEA, the UK, or Switzerland, neither one loads, and nothing beyond the server logs above is measured. Cloudflare tells the page which country the request came from, and if it cannot be checked, both stay off. You can also turn analytics off yourself, for the browser you are reading this on, with the button under "Your choices and rights" below, and that works from any country, on top of the country rule. With analytics off, neither tool loads on any page of this site from then on, and the only thing the site itself writes to your browser is the one flag that remembers the choice, which turning analytics back on removes.

PostHog records what you do on a page: which pages you open, what you click, where the pointer moves, how far you scroll, and how long the pages took to load. It also records where the visit came from and what you are reading on: the page you arrived from, any campaign tags in the link you followed, rough device and browser details, and roughly where the request came from. That location is worked out from the address: the country, the region, the city and its postal area, and the coordinates of that area, which place you in a city and no nearer. The address goes into that location and is then dropped.

PostHog also records a replay of how a page was used, which is a recording of the window as you moved through it: the pointer, the clicks, the scrolling, and the shape of the page under them. All text is masked in your browser before the recording is sent, including anything you type, so the replay shows where the words sat rather than what they said.

PostHog keeps a random identifier in your browser's local storage and sends it with what it records, so a later visit from the same browser is counted as the same visitor instead of a new one. It also keeps a few details about the visit itself in your browser's session storage, and those go when you close the tab. Neither analytics tool sets a cookie. That identifier is never tied to a name, and clearing this site's data removes it from your browser, after which the next visit starts a new one. Your browser sends what PostHog records to f.unblockedops.com, a subdomain of this site that PostHog runs for us, rather than to PostHog's own hosts.

Cloudflare Web Analytics keeps its own count of page views, page speed, and the country a request came from, as a simpler second view of the same traffic.

No advertising runs on these pages, and nothing here follows you to another site.

/book loads one more thing: the booking calendar from Cal.com, which sets a few cookies of its own and stores a couple of items in your browser so the booking works, under Cal.com's own privacy policy. Apart from that calendar and the two analytics tools, nothing else on the site loads a third party.

Booking a First Look

Bookings go through Cal.com. The form asks for a name, an email address, and a time, plus a few short questions: what is blocking you, who has to sign off and by when, whether you want a mutual NDA before the call, and how you found UnblockedOps.

Cal.com holds what you enter on its own systems and passes it to us. What the answers about how you found UnblockedOps are for is described under "How we use it" below.

The form is not the place for sensitive detail. It gets filled in before any NDA exists, so keep the answers to the shape of the problem and save the specifics for the call, or for the email that confirms it.

Send the whole thing if you have it: the security review, the questionnaire you are stuck on, the audit report. We would rather see all of it than a slice, because the size and shape of the rest is half of what we are reading for on the call. That means the whole review usually reaches us before any order exists, and it is held from the moment it arrives, under everything written on this page.

Material reaches us by reply to our confirmation email, or through an upload link we send you. No documents or attachments travel through the scheduling tool, and anything that arrives as an email attachment is moved into your folder, described under "Security" below, and deleted from the mailbox.

Engagements

Under an engagement the same kinds of material keep coming, plus what the work itself produces: the decision made about each finding, architecture and design documents, code and configuration, screenshots of consoles and findings tools.

AI tools help us run the practice, and they work on what you send us: reading a findings list, drafting a register, checking our own work against itself. Anthropic's Claude is what we use, and Anthropic processes whatever we send it for as long as a session runs. The context files those sessions work from sit in your Drive folder, as described under "Security" below.

What not to send us

Redact before you send. Logs and screenshots are the two that catch people out, because both carry personal data or secrets sitting next to the technical detail we need. If something arrives unredacted, we do not use it for the finding: we delete it, and we ask you to redact and resend.

Never send:

• Credentials, API keys, tokens, or any other secret. We take no access to your systems on the First Look or on any guided engagement. Those run on a screenshare with your engineers' hands on the keyboard, so there is no path into your environment for us to hold. If work ever needs access of any kind, it is papered separately, with its own access terms agreed before anything starts.
• Personal or health data belonging to your customers, users, or patients. We take no health data, so nothing in this work is offered to you as HIPAA-covered.
• Production data exports and database dumps.
• Employee personal data.

If any of it reaches us anyway, we delete it and tell you. A credential visible in a screenshot gets treated as exposed, which means you hear about it in time to rotate it.

How we use it

What we do with the site analytics is narrow: which pages get read, how people move between them, which ones end in a booking, where the site is slow, and where a page loses somebody. None of it tells us who you are, and we never send either tool a name.

A booking is used to prepare for the call and to follow up afterward. The answers about how you found UnblockedOps are read by hand against our own campaign list, which is how we tell which marketing is working without following you across other sites.

Material you send is used to do the work, and then to write down the shape of what we found: the tell, the root cause, the direction the fix went. Your company name, your system, and anyone who worked on it are stripped before any of it reaches our pattern library. A library full of identifiable findings would be a liability to both of us, and it would be worth less to us, not more, because what makes the next engagement faster is the shape of the problem rather than whose problem it was.

Being named is a separate matter: a case study, a testimonial, or a logo is agreed in your order, asset by asset, and never granted by this page.

Who processes it

• Google Workspace. Mail, documents, and the folder your material lives in.
• Anthropic. The AI work described above: it processes what we send it while a session is running.
• Cal.com. The booking calendar, and whatever you enter on the booking form, under its own privacy policy.
• Cloudflare. Hosting, the server logs for this website, and the page-view counts described above. It never sees engagement material.
• PostHog. The site analytics and the replay described above, and it runs the subdomain of this site that your browser sends them to. It sees visits to these pages and never engagement material.

Each one is a US company and handles what it touches under its own data-processing terms. If that list changes, this page changes with it.

Nobody outside those five companies sees your material unless the law compels it, and if that happens you hear about it from us unless we are forbidden to tell you. We do not sell your information and we do not share it for advertising.

How long we keep it

• Material you send, and the write-up or plan built from it. Twenty-four months after an engagement closes, and then it is deleted. For a First Look that never turns into an engagement, which is most of them, the same twenty-four months runs from the day we send you the write-up. We keep what you sent alongside the write-up we sent you, so that a read you later contest or we later correct can be checked against what was actually in front of us.
• What you entered on the booking form. Our own copy runs on the same clock as the material above. Cal.com keeps its copy while our account is open, with a thirty-day window to export after an account closes, under its own privacy policy.
• Session replays. Up to thirty days, the most our PostHog plan allows.
• Everything else PostHog records. It stays available to us for a year, which that plan gives us no way to shorten. Past that year the older records stop coming back in anything PostHog shows us, which is a limit on what we can see rather than a deletion we can promise you.
• The random identifier in your browser. Until you clear this site's data, as described above.

Ask us to delete your material sooner and we will, at any point. On request we clear the Drive folder, which holds the AI context files as well as the documents, along with the mail and the backups, and we confirm in writing that each one is gone. What Anthropic keeps of what we sent it during a session is the one copy we cannot clear from our own side, and a deletion that leaves a copy somewhere is not a deletion.

Your choices and rights

You can turn both analytics tools off for the browser you are reading this on, with the button below. The effect is the one described under Website visits: neither tool loads on any page of this site from then on, and the flag that remembers the choice is the only thing the site writes to your browser.

Nothing here is sold or shared, so a Do Not Track or Global Privacy Control signal from your browser changes nothing on this site.

Ask what we hold about you, ask us to correct it, ask us to delete it. Email is the whole process: [email protected]. That reaches your material and the answers on your booking form.

The site analytics work differently: nothing in them carries a name, so an email gives us nothing to match a visit against, and the way to stop them for a browser is the button above rather than a request to us. What identifies a browser there is the random identifier described above, which lives in your browser and is also recorded by PostHog against the visits your browser sent. Clearing this site's data removes it from your browser, and removes the opt-out flag with it, so the next visit starts over with the country rule applying as before and a new identifier beginning only if analytics runs, while what PostHog already holds stays attached to the old one, which carries no name.

If you sit somewhere with statutory privacy rights, such as the EU, the UK, or California, you exercise them the same way, by email.

Security

Your material lives in Google Workspace, keyed to you: one Drive folder, one mail label, one backup folder. The context files our AI tools work from live in that same Drive folder, alongside the documents they were built out of. Sharing is set on your folder and never on a parent, so nothing is reachable from above, and one person has access. That rule is held by hand rather than by the platform, because the Workspace plan carries no rule that would block a share above your folder.

Multi-factor authentication is on every account that touches client material, and full-disk encryption on every device. The practice is one person and a short list of accounts, and while that is a limit in plenty of ways, here it is the whole argument: the fewer places your material sits, the fewer places it can leak from.

If client material of yours is caught up in a security incident, you hear it from us.

Children

This is a service sold to companies. It is not directed at anyone under eighteen, and we do not knowingly collect anything from them.

International transfers

We work from the United States, and the five companies listed above are US companies. Material and booking data reaching us from Europe are handled in the United States. Which data center a processor holds its own copy in is that company's own arrangement; choosing a storage region is not a control our current Google Workspace plan offers.

Changes to this policy

When this policy changes, the date at the top changes with it. If we add anything that collects more than what is described above, or anything that recognizes you somewhere other than this site, this page changes before it goes live. Anything material, active clients hear from us directly rather than finding it here later.

Contact

UnblockedOps LLC, New York, USA. Questions and privacy requests: [email protected].