Terms of Service

These terms are an agreement between you and InTree. You accept them by continuing past the sign-in dialog (which links to this page), by creating an account, or by using the service, including the free examples that run without an account. If you use InTree on behalf of an organization, you confirm you can bind that organization to these terms.

You must be at least 16 years old and legally able to enter into this agreement.

InTree reads published biomedical literature and presents it as maps, verdicts and mechanism chains (the Explore, Verify and Trace modes), together with projects, boards, uploads, alerts, a research assistant, exports and an API for connecting external assistants. Features may change as the service develops; material changes to paid features come with notice.

InTree surfaces and summarizes published research for informational and research purposes. It is not medical advice, diagnosis or treatment, and it must not be used for clinical decision-making, patient care, emergencies, or as the basis of regulatory or safety submissions. Talk to a qualified professional for anything affecting a patient or your health.

The scientific literature is vast, and it disagrees with itself: papers contradict each other, wording is often vague or ambiguous, and study quality varies enormously. InTree's job is to connect you to what the papers actually say. It reads them, classifies each claim, weighs study quality, and labels how settled every link is, so that ambiguity is shown rather than hidden.

That reading is done by AI at a scale no one could do by hand, and it is probabilistic. A sentence can be misclassified, a relationship's direction can be wrong, evidence counts can be imperfect, coverage of the literature is partial and varies by year and source, and results may be served from a cache and not reflect the newest publications. In-product labels such as verdicts, confidence, maturity and “hypothesis” markers are themselves generated and can be wrong. Generated hypotheses and untapped links are speculative by definition and exist in no source database.

This is why every result links to its evidence. Always double-check the quoted sentences and the linked papers before you rely on, cite, or act on anything InTree shows you. Absence of a link on a map is not evidence that no relationship exists. Results and examples are illustrative; individual results vary. The service is provided as is, and descriptions of how it works are design intent, not a warranty of any particular output.

InTree is built on open, citable sources: PubMed, the Semantic Scholar Academic Graph, Reactome, WikiPathways, Open Targets, ClinicalTrials.gov and FDA FAERS. InTree is not affiliated with or endorsed by these projects or their operators. Excerpts, metadata and exports remain the work of their original authors and publishers; InTree grants you no rights to the underlying papers, and you are responsible for how you reuse exported content. Attribution for curated databases follows their licenses (for example CC-BY for Reactome).

Two source-specific cautions. FAERS entries come from spontaneous adverse-event reports and do not establish causation; they are not a substitute for a drug's label. Trial records are republished from ClinicalTrials.gov, may lag the registry, and listing a trial does not mean an intervention works.

You can sign in with Google, LinkedIn, ORCID, or an email and password. An account is created the moment you first sign in. Provider accounts sharing the same email address are linked into one InTree account, so whoever controls that email controls the account. Keep your credentials and API keys secret; you are responsible for all activity under your account. We may rate-limit, throttle, or end sessions for security at any time.

Free accounts include a monthly number of new searches; paid plans include more. One search is one new query you start; re-running a query you already ran in the same calendar month is free, and work inside an existing result (expanding nodes, opening evidence, running toolbox checks) does not use searches but has its own monthly activity ceiling to prevent abuse. Allowances reset on the first of each month (UTC) and do not roll over. When an allowance is used up, the affected feature pauses until the next month or an upgrade; nothing is ever billed as overage. The curated examples on the home page and guide run free without an account, subject to anti-abuse rate limits. We may adjust individual allowances, and staff accounts are exempt from limits.

Paid plans are subscriptions billed through Stripe, our payment processor. They renew and charge your saved payment method automatically each period until you cancel. Billing name and address are required at checkout, and your bank may require card authentication (such as 3-D Secure) to complete a purchase. Cancelling stops future renewals: you keep your plan until the end of the period you already paid for, then your account drops to Free. If a renewal charge fails, access continues for a short grace period while payment is retried; continued non-payment ends the plan.

Prices may change with notice; existing subscriptions keep their agreed rate until notified otherwise. Promotional codes may be offered and withdrawn at any time. Trials, where granted, are time-limited, require no card, and simply expire. Custom plans are separate per-account agreements.

Refunds. If you were charged in error (for example a duplicate charge) or a failure of the service prevented you from using a paid period, email support@intree.ai within 14 days of the charge and we will refund it. Otherwise, payments for billing periods that have already started are non-refundable, except where the law requires a refund.

You keep all rights to files, notes, boards and data you upload. You grant InTree the permission needed to store and process them to provide the service, including automated analysis by the AI providers we use as subprocessors. Uploaded content stays isolated to your projects: it is never added to the shared literature index, never used to answer other customers' queries, and never used to train models. Content and search history inside a shared project are visible to every member of that project; inviting someone shares the project with them. Deleting a file or project genuinely deletes its content. You can delete your entire account at any time from Settings under Privacy & data: that cancels any active subscription immediately and erases your projects, files, boards, saved papers, alerts, assistant sessions, API keys, search history and share links. It cannot be undone. Records we must keep for accounting and tax purposes are retained as described in the Privacy Policy.

Share links are viewable by anyone who has the URL, without an account, so treat anything you share as public and do not share confidential material. You can revoke a link at any time: board links from the board itself, and graph links from Settings under Privacy & data. Revoking deletes the shared copy and the URL stops working, though anything a viewer already saved or exported stays with them. Links created while signed out belong to no account and cannot be revoked. Exports (CSV, citations, BibTeX) contain excerpts and metadata from third-party publications, covered by section 5.

API keys let external assistants and scripts use InTree as your account. A key is a full-account credential: anything done with it counts as you and consumes your plan's allowance. Keep keys out of shared URLs and public code, and revoke any key you no longer use. Automated access must stay within your plan's limits; bulk scraping of the service or systematic redistribution of its data is not permitted.

Use InTree lawfully. Do not abuse, overload, probe, or attempt unauthorized access to the service or other users' data; do not reverse-engineer it; do not resell it or misrepresent its output as human expert curation. We may throttle, suspend, or terminate accounts that violate these terms, and we may remove content that we must remove by law.

InTree is provided as is and as available, without warranties of any kind, including fitness for a particular purpose and non-infringement. We depend on third-party sources and APIs and do not warrant uninterrupted service or the accuracy of any output (see section 4). To the fullest extent the law allows, InTree's total liability for any claims under these terms is limited to the amount you paid us in the twelve months before the claim; we are not liable for indirect, incidental, or consequential damages, including decisions made in reliance on results.

We may update these terms. When we make material changes we bump the version above and show a notice in the product asking you to review and accept; continuing to use InTree after changes constitutes acceptance. The current version always lives at this address.

Questions, privacy requests and refunds: support@intree.ai. Account deletion is self-serve in Settings under Privacy & data.