Justin McKelvey
Ran snowboard shops. Taught myself to code. Now I build AI that does real work.
Is Claude HIPAA Compliant? Which Plans Get a BAA (2026)
Quick Answer
No, not on Free, Pro, Max, Ultra or Team. No model is HIPAA compliant by itself; what counts is a signed Business Associate Agreement (BAA). Anthropic signs one only for the Claude API, Claude in Microsoft Foundry and Claude Enterprise (self-serve or sales-assisted, 20-seat minimum), and Enterprise is covered only after the Primary Owner turns HIPAA on and accepts the BAA. Even then, Claude Code and Cowork are covered in local mode only, and the Console, Cowork in the cloud and beta features like Claude in Office stay outside.
Checked against Anthropic's BAA article, its HIPAA-ready Enterprise help page and claude.com/pricing on November 8, 2026 · Author: Justin McKelvey, fractional CTO and AI consultant, 15 years in software, 50+ products shipped
Every few weeks someone asks me a version of the same question, usually a little sheepishly, usually after the fact: our front desk has been using Claude to draft patient follow-ups. Is that a problem?
The internet answers this badly. Search it and you'll get a wall of confident yes/no verdicts, most of them written by companies selling a "HIPAA-compliant AI wrapper," most of them skipping the one distinction that actually decides your answer. So here it is first, before anything else.
No model is HIPAA compliant. A contract is.
HIPAA compliance is not a property that software has. There is no certification body that inspects a language model and stamps it. What exists is a Business Associate Agreement: a contract in which a vendor handling protected health information on your behalf accepts legal obligations for safeguarding it.
So "is Claude HIPAA compliant?" is the wrong shape of question. The right one is: will this vendor sign a BAA covering the specific product I am about to paste a patient's name into? That question has a precise, checkable answer, and it changes tier by tier within the same brand.
Which is why the yes/no articles are so useless. The answer for Claude is genuinely both. It depends entirely on which Claude you're using, and almost certainly not the one you're using.
What Anthropic will and won't sign for
Anthropic's published BAA documentation draws the line in two places:
| Claude product | Typical price | BAA available? |
|---|---|---|
| Claude Free | $0 | No |
| Claude Pro | $20/month ($17/month billed annually) | No |
| Claude Max | $100/month | No |
| Claude Ultra (called Max 20x until November 1, 2026) | $200/month | No |
| Claude Team | Standard seat $20/month annual ($25 monthly), premium seat $100 ($125 monthly), teams of 2 to 150 | No |
| Claude Cowork in the cloud | Bundled with paid plans | No (local mode can be covered on HIPAA-ready Enterprise once configured) |
| Claude Enterprise (self-serve) | $20/seat/month annual, 20-seat minimum, plus usage at API rates | Yes, once the Primary Owner activates it (changed October 2026) |
| Claude Enterprise (sales-assisted) | $20/seat/month annual plus usage, through sales | Yes: must be activated |
| Claude API (first-party) | Metered, pay-as-you-go | Yes |
Two details in that table do real damage if you miss them.
Update, October 10, 2026: self-serve Enterprise now qualifies. When this page launched in September, Anthropic tied the BAA to the sales-assisted Enterprise plan, and buying Enterprise with a card did not get you one. Anthropic's BAA article, updated this week, now says the BAA covers "Enterprise plans" and that Enterprise Primary Owners "can accept the BAA directly" in settings, and its pricing page marks a HIPAA-ready offering on both Enterprise columns (self-serve and sales-assisted) and not on Team. Both flavors cost the same, $20 a seat a month billed annually plus usage at API rates; what still differs is invoicing, net terms and support. The published number and the Team-vs-Enterprise decision are on the Claude Enterprise pricing page.
The coverage still stops at the feature line. Even on a covered Enterprise organization, Anthropic lists Claude Console, Cowork in the cloud, and beta features (Claude in Office, Claude Design, Claude Slides, Claude Docs) as outside the BAA. Claude Code and Cowork in local mode can be covered, but only after the Primary Owner applies the HIPAA configuration to both. So the practical question for a practice is no longer which Enterprise you bought; it is which features your staff are allowed to open.
And HIPAA coverage on Enterprise is not automatic once you have it. Anthropic requires the organisation's Primary Owner to actively switch on HIPAA compliance in the workspace's Data and privacy settings and accept the BAA. A standard Enterprise organisation is not covered until somebody does that. If you're on Enterprise today and nobody can tell you who accepted the BAA or when, assume it hasn't happened.
One more piece of housekeeping worth checking if you've been around a while: Anthropic has stated that BAAs signed before 2 December 2025 cover API usage only and do not extend to the HIPAA-ready Enterprise plan. An old agreement in a drawer may not cover what you think it covers.
If you're choosing between vendors rather than checking one, all four (Claude, ChatGPT, Gemini and Copilot) are lined up in one table, with what the covered path costs a 10-person practice, on HIPAA compliant AI.
How OpenAI and Microsoft draw the same line
OpenAI draws the line in the same conceptual place, and the tiers that make it across are the enterprise-and-API ones:
- Covered by a BAA: ChatGPT Enterprise (the sales-managed one), OpenAI's healthcare-specific enterprise products, and the API Platform, and notably the API does not require an enterprise agreement to get a BAA, which matters enormously for small practices.
- Not covered: Free, Go, Plus, Pro, and ChatGPT Business (the self-serve team plan that used to be called Team).
The full OpenAI answer, including the free ChatGPT for Clinicians BAA and the features OpenAI leaves outside even a signed one, is on Is ChatGPT HIPAA compliant?
The naming here is genuinely hazardous. OpenAI ships a consumer-facing health product whose name contains the word "health" and which is not BAA-covered, sitting in a catalogue alongside enterprise healthcare products that are. If your compliance reasoning is "well, it's called Health, so presumably…", stop there. The name is marketing; the BAA is the fact. If you're weighing the two vendors more broadly, Anthropic vs OpenAI for business covers the rest of the comparison.
And a third vendor breaks the pattern entirely, which is worth knowing before you buy anything. Microsoft offers its HIPAA BAA by default through the Data Protection Addendum every commercial Microsoft 365 customer already accepted, no sales call, no signature, no extra fee, and it names Microsoft 365 Copilot on its published in-scope list. If your practice already runs on Microsoft 365, you may be holding a signed BAA covering a tool your staff uses daily, while shopping for an enterprise contract you don't need. The exposure there sits somewhere else entirely: is Microsoft Copilot HIPAA compliant has the two gaps that actually matter.
OpenAI publishes no price for ChatGPT Enterprise, so on the ChatGPT side the BAA route starts with a sales call. What the Enterprise contract includes, and how OpenAI bills it, is in ChatGPT Enterprise pricing.
Is Claude Code HIPAA compliant? Cowork, the API and the cloud platforms
Claude Code and Cowork can be covered, but only in local mode, only on a HIPAA-ready Enterprise organization, and only after the Primary Owner applies the HIPAA configuration. Anthropic's help page on HIPAA-ready Enterprise plans says that once the configuration is on, "Your BAA covers Claude Code (local mode) and Cowork (local mode)." Cloud features, including Claude Code cloud sessions, Claude Code Remote Control and Cowork in the cloud, aren't available at all in that setup. Before the configuration is applied, Cowork works but isn't covered, and Claude Code is covered only with zero data retention, which Anthropic limits to qualified accounts.
Turning HIPAA on is a one-way door. Only the Primary Owner can do it, the BAA is a click-to-accept step inside the same flow, and Anthropic says the change can't be undone from organization settings. Decide who that person is, and get your privacy officer's sign-off, before anyone clicks.
The API has its own carve-outs. Anthropic's BAA article lists the Batch API, the Files API, the Skills API, code execution and web fetch as not covered and not available to HIPAA-ready API users. Anything you pass on to a third party, through a connector or an external MCP server, isn't covered by Anthropic's BAA either. Covered models run with 30-day data retention, and zero data retention needs Anthropic's approval.
On the cloud platforms, the door you use decides it. Anthropic lists Claude through Microsoft Foundry as eligible under its BAA, though some features, such as the Compliance API, aren't available there. Claude Code run through Amazon Bedrock or Google Cloud is "available but not covered" under Anthropic's BAA, and local Claude Code is covered only when it connects directly to Anthropic. If you run Claude on AWS or Google Cloud, the agreement that matters is your BAA with that cloud provider, so check its list of covered services before any patient data goes in.
The plans that trip everyone up: Claude Team and ChatGPT Business
If you take one thing from this page, take this one, because it is the mistake I see most and it is an expensive one.
Claude Team and ChatGPT Business are not HIPAA-eligible.
These are precisely the plans a five-person practice talks itself into. They cost real money. They say team and business on the label. They come with admin controls, shared workspaces, and a contractual promise not to train on your data, and that last one is what does the damage, because "they don't train on our data" feels like the compliance box being ticked.
It isn't. Training policy and BAA coverage are two unrelated things. You can be on a plan that will never train on your inputs and still have no agreement in place governing protected health information, which means every PHI-containing prompt is a disclosure to a vendor who never accepted responsibility for it. Upgrading from Pro to Team moved you into better admin tooling and left your HIPAA position exactly where it was. (I've written separately about what the Claude Team plan actually buys you, it's a good plan. It is not this.)
This matters most for rural clinics and community health centers, because Anthropic's nonprofit program lists Critical Access Hospitals, Rural Emergency Hospitals, Federally Qualified Health Centers and Rural Health Clinics as eligible for Claude Team at $8 a user a month. The discount changes the price, not the paperwork. What the program covers is in Claude for nonprofits.
Signing a BAA doesn't make you compliant either
The other half of the trap sits on the far side of the contract. A BAA is a shared-responsibility document. It covers the vendor's obligations. It says nothing about yours.
Still entirely on you, after signature:
- Access control: who in the practice can open the workspace, and what happens to that access the week someone leaves.
- Minimum necessary: whether the prompt needed the patient's full name and date of birth to do the job, or whether that was just how the front desk happened to paste it.
- Staff training: because the actual breach vector is somebody using their personal ChatGPT on their phone at lunch, not your carefully configured enterprise workspace.
- Breach detection and notification: you need to be able to tell that something went into the wrong tool, and you're on the clock once you know.
- Documentation: a written policy naming which tools are approved for what. If it isn't written down, it doesn't exist in an audit.
A signed BAA is a starting gun, not a finish line.
The cheapest legal answer is usually "don't send PHI at all"
Here's the part the compliance-wrapper vendors would rather not lead with. A large share of what a practice actually wants AI for does not require protected health information in the first place.
Drafting a recall campaign. Explaining what a benefits breakdown means in plain English. Writing a procedure explainer for the website. Turning a messy claims narrative into a reusable template. Rewriting your no-show policy so it sounds like a human wrote it. None of those need a name, a date of birth, or a chart number to work, and the moment they don't, the entire HIPAA question stops applying to that workflow.
The micro-action, and you can do it today: open whatever AI tool your practice is already using, look at the last ten prompts your team ran, and mark each one needed PHI or didn't. In every practice I've done this with, the second pile is much bigger than anyone expected. That exercise takes fifteen minutes and it tells you whether you have a contract problem or a habit problem, and they have very different price tags.
If the workflow genuinely does need PHI, price the API route with a BAA next to the Enterprise quote before you sign anything. For a small practice running modest volume, metered API usage under a BAA is frequently far cheaper than a per-seat enterprise agreement sized for a hospital, and it's the option nobody's sales team is incentivised to mention. For how the consumer tiers price out by comparison, Claude's pricing breakdown has the full ladder.
Law firms ask the same question with a different rulebook: client confidentiality instead of PHI, and the bar instead of HHS. The plan answer is similar, and it's laid out for a small practice in Claude for lawyers.
For a practice that wants this planned and set up rather than worked out in-house, AI consulting for healthcare practices has every price and the plan check I run first.
OpenAI's closest equivalent for a single provider is free, and it covers one person. The details are on ChatGPT for Clinicians.
What I tell owners
Almost nobody who asks me this question has a HIPAA problem. They have an unwritten-rules problem that a HIPAA question finally surfaced.
Staff adopted a tool because it made a genuinely tedious job faster, nobody said not to, and eighteen months later there's no policy, no approved-tools list, no idea which subscriptions are personal versus practice-owned, and no one who can say with confidence what has been pasted where. The compliance exposure is real, but it's a symptom. The disease is that AI arrived in the practice the way a stray cat arrives at a back door, incrementally, affectionately, and with nobody making a decision.
The fix is unglamorous and it is mostly not a purchase: decide which two or three workflows AI is actually for, write down which tool is approved for each and what may never be pasted into any of them, and de-identify by default so the expensive contract question only applies to the small number of workflows that truly need it.
If you'd rather not work out that order yourself, the AI Readiness Assessment ($2,500 flat, two weeks, a written roadmap, fee credited against a build within 90 days) is exactly this exercise done properly: which processes, in what order, on which tools, with the governance written down instead of assumed. The do-it-yourself version is the CFSB Playbook ($397). And if you want to talk through a specific practice, that's what strategy calls are for. Practice-specific workflow examples live in AI for dental practices.
One last thing, and I'd rather say it plainly than hedge it: this is a technology explainer written by a fractional CTO who implements these systems, not legal advice. HIPAA enforcement turns on specifics, your data, your workflows, your risk analysis. Use this to ask your compliance counsel a much sharper question than "is ChatGPT okay?", which is a question nobody can answer well. "Which of our AI tools are covered by a signed BAA, and who accepted it?" is the one worth asking.
Related: the same terms-first read applied to a vendor whose policy points the other way, data stored in the People's Republic of China and used for training unless you opt out, is on Is DeepSeek safe?
Get the Free AI Content Toolkit
The exact system I use to turn one idea into a month of content: atomization framework, voice template, prompt library, weekly system.
Frequently Asked Questions
- Is Claude HIPAA compliant?
- Not by itself, and no AI model is. HIPAA compliance isn't a property a model has, it's a contract, a Business Associate Agreement (BAA), that a vendor signs committing to handle protected health information under the rules. Anthropic will sign one, but only for its commercial HIPAA-ready services: the first-party Claude API, Claude in Microsoft Foundry, and Claude Enterprise plans, where the organization's Primary Owner turns HIPAA on and accepts the BAA in settings (as of October 2026, that includes Enterprise bought self-serve). Claude Free, Pro, Max, Ultra and Team are outside that coverage, and so are Cowork in the cloud and beta features like Claude in Office, even on Enterprise. So if you're logging into claude.ai on a Pro subscription, the honest answer is no, and no amount of careful prompting changes it.
- Is Claude Code HIPAA compliant?
- Only in local mode on a HIPAA-ready Claude Enterprise organization, after the Primary Owner applies the HIPAA configuration. As of November 2026, Anthropic says the BAA then covers Claude Code (local mode) and Cowork (local mode), while Claude Code cloud sessions and Remote Control aren't available. Claude Code run through Amazon Bedrock or Google Cloud is not covered under Anthropic's BAA, and local Claude Code is covered only when it connects directly to Anthropic. On Pro, Max, Ultra or Team there is no BAA at all.
- Is Claude Cowork HIPAA compliant?
- Cowork in local mode can be covered on a HIPAA-ready Claude Enterprise organization once the Primary Owner applies the HIPAA configuration. Cowork in the cloud is excluded from Anthropic's BAA, and before the configuration is applied Cowork is available but not covered. On Pro, Max, Ultra and Team, Cowork has no BAA, so no patient information goes into it.
- Can I use Claude Pro or ChatGPT Plus with patient data?
- No. Those are consumer tiers and neither vendor will sign a BAA covering them, which means putting protected health information into one is a disclosure to a party with no agreement in place. That's the part people get wrong, they assume the risk is that the model 'learns' their data, so they turn off training and consider it handled. Training settings aren't the control that matters here. The absence of a signed BAA is the problem, and you can't fix it from inside the settings menu.
- Is Claude Team or ChatGPT Business HIPAA compliant?
- No, and these two are the expensive mistake. They're the plans a small practice naturally buys, they cost real money, they say 'team' and 'business' on the tin, and they come with admin controls and no-training contracts that feel like compliance. Neither is BAA-eligible. Anthropic excludes Claude Team; OpenAI excludes self-serve ChatGPT Business. Paying more did not move you into coverage. It moved you into a plan with better admin settings and exactly the same HIPAA answer as the free tier.
- Does signing a BAA make my practice HIPAA compliant?
- No, and this is where a lot of small practices stop reading too early. A BAA is a shared-responsibility document: it covers the vendor's obligations for the data you send. Everything on your side is still yours, who has access, how staff are trained, what goes into a prompt in the first place, how you'd detect and report a breach, and whether the workflow needed the patient's name at all. I've watched practices treat a signed BAA as a finish line when it's closer to a starting gun.
- What's the cheapest way to actually use AI in a practice legally?
- Usually: don't send PHI at all. A surprising share of what practices want AI for, drafting recall messages, summarising an insurance policy, writing a procedure explainer, cleaning up a claims narrative template, doesn't need a patient's identity to work. De-identify first and most of the compliance question evaporates along with the cost of an enterprise contract. If the workflow genuinely needs PHI, then you're in BAA territory and you should price the API route alongside Enterprise, because for a small practice the API with a BAA is often dramatically cheaper than a per-seat enterprise agreement.
More on AI for Business
ChatGPT Business vs Plus (2026): Which One a Small Business Should Actually Pay For
ChatGPT Plus vs Business, as of September 2026: Plus is $20 a month for one person and OpenAI may train on your chats unless you opt out; Business is $20 a seat billed annually or $25 monthly, two seats minimum, with no training by default, admin controls, shared projects for groups, and the Company Knowledge plugin. The side-by-side from OpenAI's pages, and three owner scenarios with a verdict each.
ChatGPT for Customer Service (2026): What Works, What Breaks, and What It Costs a Small Business
ChatGPT for customer service, as of September 2026: it drafts replies, answers policy questions from your own documents, and triages an inbox for $20 to $25 a seat a month on ChatGPT Business. It can't look up an order, issue a refund, or remember a customer unless you connect the system that holds them. The honest capability line, the plan you need, a setup in six steps, the failure modes, and the point where a real agent takes over.
ChatGPT Pulse (2026): What It Was, Why OpenAI Retired It, and How to Get the Morning Brief for Your Business
ChatGPT Pulse, the daily research cards OpenAI previewed for Pro users in September 2025, was sunset on June 17, 2026 with a 14-day wind-down. What it did, which plans ever had it, why scheduled tasks replaced it, and how an owner rebuilds the morning brief on a Plus or Business seat without handing ChatGPT the wrong things.
Copilot Cowork (2026): What It Does, Who Gets It, and What It Really Costs Per Month
Copilot Cowork is Microsoft's agent that runs long, multi-step tasks across Outlook, Teams, Word, Excel and PowerPoint. It needs a Microsoft 365 Copilot license and then bills each task in Copilot Credits at $0.01 each, with no published price per task. It's off until an admin turns it on, and since Ignite (Nov 17, 2026) Copilot Chat can offer to hand your request to it.
Written by
Justin McKelvey
I spent 12 years running snowboard and mountain-bike shops in Colorado, then taught myself to code. Since then: 50+ products shipped, Head of Product at a national sports platform, CTO at an AI startup. Now I put AI to work in real businesses in Austin, TX and remote, and write down what actually works.
Work with meWant more help? Get the free toolkit · Book a strategy call
Before you go
Score your AI readiness in 3 minutes.
Free, no call. 30 yes/no questions show where AI actually pays off in your business (and where it doesn't yet).
Get my readiness score →