Terms of service

Helm is a personal medical records portal with an AI-driven research assistant. It helps you organize your own health data, surface patterns Captain (an AI) identifies in that data, and prepare for conversations with your healthcare providers.

Helm is positioned as a general wellness product under FDA guidance — a software tool that supports organizing personal health information and discussing health with providers. It is not a medical device.

Helm is not:

• A medical practice, clinic, or healthcare provider
• A telemedicine service
• A diagnostic tool — Captain produces theories, not diagnoses
• A replacement for professional medical advice or care
• A pharmacy or prescriber
• A health insurer or claims processor
• A medical device under FDA regulation
• A HIPAA-covered entity (see Privacy Policy §11 for our HIPAA-grade architecture vs. covered-entity status)
• An emergency service — if you're in crisis, call 911 or 988 (US suicide and crisis lifeline) or visit an emergency room

Helm is early-access software. By using Helm, you understand and agree that:

• Features may change, break, or be removed without notice
• The product may be temporarily unavailable
• AI behavior (Captain's output style, accuracy, coverage) may shift as we update the underlying models or prompts
• Bugs are possible. Some may affect how data is displayed or organized — though we take care to avoid bugs that affect data integrity
• Paid memberships are available. Plan prices, trial terms, included usage, and any applicable charges are shown before you purchase (see §8).
• You should not rely on Helm for any urgent or critical health decision. Always go through a licensed provider for anything time-sensitive or consequential.

Captain is Helm's AI assistant. It uses third-party AI technology provided by Anthropic and accessed through AWS Bedrock under our AWS Business Associate Agreement. It reads the data in your account and generates structured output: case narratives, ranked diagnostic theories, recommended diagnostic tests, ruled-out logs, research references, visit-prep packets, and conversational responses.

Captain can be wrong. It can miss things, it can over-weight things, it can produce plausible-sounding output that's factually incorrect (“hallucinate”), and it can fail to recognize emergencies or red-flag symptoms. It is a tool to help you think — it is not a doctor, it does not have your best interests in the way a person does, and it cannot replace clinical judgment from a licensed healthcare provider who has examined you.

Always validate Captain's output with a licensed provider before acting on it. Specifically:

• Don't start, stop, or change medications based on Captain's output alone
• Don't skip diagnostic workups your doctor wants based on Captain's ruled-out log
• Don't delay seeking emergency care because Captain didn't flag something as urgent
• Don't use Captain to self-diagnose serious or progressive conditions without provider involvement

The research references Captain surfaces (PubMed papers, ClinicalTrials.gov listings, published authors) are informational pointers — we don't verify the scientific validity, current accuracy, or appropriateness of any specific paper, study, trial, or specialist for your individual case.

Captain may also call tools (structured actions) that modify your account — e.g., adding an appointment, logging a symptom, editing your case narrative. We surface every tool call as a labeled chip in the chat so you can see exactly what changed. You can undo many tool actions via the in-app controls. If something looks wrong, fix it manually or email us.

You agree to:

• Provide accurate information when you create your account
• Keep your password secure — anyone with your credentials can read everything in your account. Enable MFA in Settings if you want a second layer of protection.
• Use Helm for your own health data, or with permission for data of someone you legitimately care for (e.g., a parent for a minor child, or a caregiver with explicit consent)
• Be at least 18 years old to create your own account. A minor's health information may be managed only by their parent or legal guardian through that adult's account (the Family feature) — a minor may not create their own account.
• Notify us if you suspect unauthorized access to your account

You're responsible for what you upload. Don't upload data you don't have the right to share with us for processing.

You will not:

• Attempt to access other users' data, accounts, or session tokens
• Reverse-engineer, decompile, or attempt to extract the source code of Helm, its API, or its AI prompts
• Use automated tools (scrapers, bots, crawlers) against Helm's interfaces in ways that could degrade service for other users
• Submit prompt-injection attacks intended to coerce Captain into revealing system prompts, generating harmful content, or bypassing safety measures
• Use Helm to harass, defame, or endanger anyone — including by uploading data about a third party without their consent
• Use Helm to coordinate or facilitate the avoidance of legitimate medical care for a third party (e.g., a child you have custody of)
• Use Helm in connection with any illegal activity, including violations of state or federal drug, prescription, or insurance-fraud law
• Resell, sublicense, or otherwise commercialize access to Helm without our written consent
• Interfere with the integrity or operation of Helm, including by exploiting rate limits, abusing free-tier features, or attempting to circumvent access controls

We reserve the right to suspend or terminate accounts that violate this section, with reasonable notice where possible.

You control sharing. When you generate a doctor share link or invite a caregiver, anyone who accesses that link gains read-only visibility into the data we've included in the share payload. You are responsible for who you share with and for any consequences of that sharing.

Helm doesn't verify the identity of doctor-share-link recipients — the bearer token is the credential. Don't forward a share URL to anyone you don't intend to give access to.

See the Privacy Policy (/privacy) §7 for the full mechanics of doctor share links, caregiver access, and record-share tracking.

Managing someone else's information (caregivers & dependents)

Helm lets you manage another person's case — as a caregiver invited by an adult patient, or as the parent or legal guardian of a dependent (such as a minor child) who doesn't have their own login. When you do, you represent and warrant that you are legally authorized to access and manage that person's health information — for example, that the patient invited you, or that you are the parent, legal guardian, healthcare proxy, or other authorized representative of the person whose information you add or manage.

Helm does not independently verify that authority and is not responsible for determining it. You are solely responsible for ensuring you have the right to access and manage the information and for complying with any custody orders, guardianship terms, or other legal arrangements that apply. You agree to stop accessing a person's information — and to remove the relationship — if your authority ends or is revoked. The patient (or, for a dependent, another authorized guardian) may revoke access at any time, and all access is logged.

We are not a HIPAA covered entity, and managing another person's information through Helm does not create a provider relationship. If you are not authorized to access someone's health information, do not add it to Helm.

We make a small number of specific claims about Helm in our marketing and policies:

• That Helm uses HIPAA-grade architecture — encryption at rest + in transit, access controls, audit logging, AWS BAA coverage, data portability + deletion. We do not claim to be a HIPAA-audited covered entity (we're not).
• That we do not sell your data, do not share it with advertisers, and do not use it to train AI models.
• That AWS Bedrock (and underlying Anthropic inference) is covered by our BAA and does not retain inference data beyond the request.
• That our Privacy Policy identifies the service providers that handle data for Helm: AWS for infrastructure and AI services; Stripe for checkout, subscriptions, and payment processing; and Sinch for sending and receiving fax documents when you use the fax feature, including the health information those documents contain.

These representations are made truthfully under FTC Act §5. If we change any of them, this document and the Privacy Policy will be updated, our internal LEGAL_DOCS_VERSION will be bumped, and you'll be re-prompted via the /welcome consent screen on your next sign-in. If you believe we've made a representation that's no longer accurate, please email us — we take this seriously.

Helm offers paid subscription tiers through Stripe. When you start a plan that includes a trial, Stripe collects a payment method at checkout, Helm starts the trial period shown at checkout, and Stripe automatically charges the selected plan at the end of that trial unless you cancel first.

• Current launch trials are 7 days. You are not charged during the trial period.
• Plan prices, included usage, feature limits, add-ons, taxes, and any overage or catch-up pack charges are shown in Stripe Checkout or in the plan management flow before purchase.
• You can manage payment methods, invoices, cancellation, and renewal through Settings → Plan or the Stripe Customer Portal. Cancellation stops future renewals; it does not automatically delete your Helm account or health data.
• Refunds are not guaranteed unless required by law, but you can contact billing@helm.org if something looks wrong.

You can close your account at any time from Settings → Privacy & data → Delete account. After a 90-day grace period, your data is permanently deleted (see Privacy Policy §13 for retention details). We don't charge a fee for account closure.

We aim for high availability but don't guarantee it. Helm may be temporarily unavailable for maintenance, due to upstream provider issues (AWS, Anthropic), or for reasons outside our control. We don't provide SLAs or uptime guarantees in this version.

Captain's output quality depends on the upstream AI model. We can't guarantee specific output behavior; model updates from Anthropic may change response style, accuracy, or coverage over time.

We may temporarily disable specific features (voice mode, AI re-reads, etc.) if a downstream service fails or if usage patterns require it. We'll communicate via in-app messaging when we do.

To the extent allowed by law:

• Helm and its operators are not liable for medical decisions you (or others) make based on Captain's output. You are responsible for validating with a licensed provider before acting.
• Helm is not liable for misdiagnosis, missed diagnosis, delayed treatment, adverse drug events, or any other clinical outcome attributable to Captain's output or your use of Helm.
• Helm is not liable for lost data due to your account closure, account compromise (via your own credentials), or upstream service outages.
• Helm's total aggregate liability for any claim arising from your use of the service is limited to the greater of (a) the fees you've paid to us in the 12 months preceding the claim, or (b) $100 USD.
• Helm is provided “as is” without warranties, express or implied, including warranties of merchantability, fitness for any particular purpose, accuracy, or non-infringement.
• Helm specifically disclaims any warranty regarding the accuracy, completeness, or appropriateness of Captain's AI-generated output.
• Some jurisdictions don't allow certain limitations of liability. If you're in one of those jurisdictions, the above may not fully apply, but applies to the maximum extent allowed.

You agree to indemnify and hold harmless Helm, its operator (Helmworks LLC), and any future contractors, employees, or affiliates from any claim, damage, or expense (including reasonable attorneys' fees) arising out of:

• Your violation of these terms or our Privacy Policy
• Your use of Helm to upload data you didn't have the right to upload
• Decisions made by you or anyone you shared data with based on Captain's output
• Your sharing of data via Helm with third parties

This section does not require you to indemnify us for our own willful misconduct or gross negligence.

You retain all rights to the medical data you upload. We don't claim ownership of your records, your symptom logs, or any other content you provide.

We retain all rights to the Helm application code, branding, UI design, AI prompts, and methodology. You may not copy, redistribute, resell, or create derivative works of the application or its components without our written consent.

Captain's output (reflections, theories, narratives, etc.) is generated for your use. You may use, share, print, forward, or repurpose it freely as it relates to your own healthcare. You may not use Captain's output to train other AI models or to build a competing service.

HELM, TAKE THE HELM and CAPTAIN are trademarks of Helmworks LLC.

These terms are governed by the laws of the United States and the State of California, without regard to conflict-of- laws principles.

Informal resolution first. Before filing any formal claim, you agree to contact us at hello@helm.org with a description of the dispute and what you'd like to resolve. We'll respond within 30 days and try to work it out in good faith. Most issues can be resolved here.

Jurisdiction. If informal resolution doesn't work, any disputes will be resolved in the state or federal courts located in San Francisco County, California. Both parties consent to personal jurisdiction there.

Right to file a regulatory complaint. Nothing in these terms prevents you from filing a complaint with a government regulator (FTC, state attorney general, etc.) at any time. Some state privacy laws (e.g., Washington MHMDA) also give you a private right to sue regardless of these terms.

We may update these terms as the product evolves. Material changes trigger:

• In-app re-consent via the /welcome screen — we bump our internal legal-docs version constant, which forces every signed-in user back through the consent gate on their next session
• An email to your account address explaining what changed

Continued use after a change (after re-consenting) constitutes acceptance of the updated terms.

If any provision of these terms is held invalid or unenforceable, the remaining provisions stay in effect. These terms (together with the Privacy Policy) constitute the entire agreement between you and Helm regarding your use of the service.

Questions about these terms: hello@helm.org
Privacy-specific: privacy@helm.org