Terms and privacy

Users are granted a personal, non-exclusive, non-transferable, revocable, and non-sublicensable license to interactively explore pharmacokinetic and pharmacodynamic scenarios, evaluate multi-drug interaction hypotheses, and examine biophysical translation mechanics through the web graphical user interface.

You agree not to use, deploy, or facilitate any automated system, script, bot, spider, scraper, crawler, or programmatic routine to access, monitor, query, harvest, extract, or copy the catalog of compounds, kinetic parameters, mechanistic rule definitions, or simulation outputs.

• Direct programmatic invocation of backend simulation or session endpoints outside the authenticated official web client is strictly prohibited.
• Circumventing, forging, or tampering with cryptographic session signatures, rate-limiting headers, or client validation tokens constitutes a material breach.
• Bulk automated querying intended to reconstruct the underlying differential equation matrices, parameter libraries, or clinical translation tables is actionable at law.

Users shall not decompile, reverse engineer, disassemble, decrypt, extract, or derive the proprietary mathematical representations, kinetic parameter estimations, population distributions, or clinical rule heuristics embedded within the Affinix computational engine.

Publishing comparative benchmarking studies, competitive performance assessments, or systematic accuracy evaluations of the Affinix computational solver without prior written methodology verification and express authorization from the publisher is strictly prohibited.

Statutory database protection attaches to the entire structured collection, including but not limited to: 1.1K+ curated pharmacological compounds; individual two-compartment ODE kinetic parameter vectors (CL, V1, V2, Q, ka, fu); clinical trial-benchmarked half-lives and Tmax ranges; qualitative and quantitative CYP/UGT/transporter metabolic fractionation schemas (fm); time-dependent mechanism-based inactivation constants (kinact, Ki); and the deterministic biophysical clinical translation rule algorithms.

The creation, verification, and continuous maintenance of this pharmacometric compendium have required hundreds of hours of forensic literature review, peer-reviewed clinical data extraction, cross-referencing across EMA/FDA regulatory review dossiers, in-silico reconciliation of contradictory pharmacokinetic literature, and bespoke mathematical ODE tuning.

In accordance with Article L342-1 of the French Intellectual Property Code and Article 7 of Directive 96/9/EC, the following acts are strictly prohibited without the express prior written consent of the database producer:

• The permanent or temporary transfer of all or a substantial part, evaluated qualitatively or quantitatively, of the contents of the database to another medium by any means or in any form whatsoever (extraction).
• The making available to the public of all or a substantial part, evaluated qualitatively or quantitatively, of the contents of the database by the distribution of copies, by renting, by on-line or other forms of transmission (re-utilization).
• The repeated and systematic extraction or re-utilization of insubstantial parts of the contents of the database implying acts which conflict with a normal exploitation of that database or which unreasonably prejudice the legitimate interests of the maker of the database.

Infringements of the database producer rights guaranteed by Article L342-1 of the French Intellectual Property Code are punishable under Article L343-4 by three years imprisonment and a fine of up to €300,000, alongside civil damages, preliminary and permanent injunctive relief, and the seizure of infringing datasets or downstream software products.

No entity, automated system, corporate organization, or research laboratory may use, ingest, tokenize, fine-tune, distill, evaluate, or embed any portion of the Affinix compendium, scientific text, mechanistic logic, or API simulation outputs to train, fine-tune, or distill weights for foundational artificial intelligence models, large language models (LLMs), or deep neural networks without prior written commercial licensing. In accordance with platform machine-readable Content-Signal directives (search=yes, ai-train=no, use=reference), ephemeral read access for interactive user citations and real-time search reference via standard RFC 9110 content negotiation is permitted, provided no data is retained for model training or synthetic distillation.

This reservation is formally implemented in machine-readable formats across the platform infrastructure:

• Structured machine-readable Content-Signal directives and HTTP metadata headers (X-Robots-Tag: noai, noimageai) reserving all rights against artificial intelligence training and model distillation.
• Structured programmatic declarations delivered via /.well-known/ai-catalog.json and /llms.txt specifying commercial reservation and reference-only usage.
• Perimeter-level validation, cryptographic session token verification, request payload caps, and automated query rate limiting on computational simulation endpoints.

Generating simulated scenarios programmatically to capture synthetic output pairs (e.g. drug input vector → concentration curves/clinical alerts) for the purpose of distilling, training, fine-tuning, or evaluating machine learning models is an express violation of this reservation and constitutes unlawful extraction under Article L342-1 of the French Intellectual Property Code.

All concentration-time curves, receptor occupancy estimates, metabolic clearance adjustments, population envelopes, and clinical alerts generated by Affinix are simplified mathematical approximations computed from deterministic ordinary differential equations (ODEs), standard two-compartment assumptions, and literature-derived parameters. They reflect idealized in-silico models rather than measured biological reality.

Affinix does not constitute a Clinical Decision Support System (CDSS) or Software as a Medical Device (SaMD). It must never be relied upon by physicians, pharmacists, nurses, clinical researchers, or patients to determine, adjust, withhold, or discontinue dosages, therapeutic regimens, or drug combinations for any individual human patient.

Real-world clinical pharmacokinetics and pharmacodynamics are governed by numerous unmodeled patient-specific factors, including but not limited to: renal and hepatic hemodynamic shifts, uncharacterized genetic polymorphisms, variable GI motility and microbiome metabolism, organ blood flow alterations, acute phase protein binding changes, and patient adherence. The simulation engine cannot anticipate idiosyncratic or unmodeled biological responses.

To the maximum extent permitted by applicable French and European law, the publisher disclaims all liability for any direct, indirect, incidental, consequential, special, or exemplary damages, medical adverse events, bodily injury, or loss of life arising out of or in connection with any reliance placed upon the calculations, curves, or alerts generated by the platform.

The Affinix website, software platform, and computational services are edited and published by: Charles Duprat, registered under French business registry SIREN 881 003 255, residing in Marseille, France. Editorial direction and technical management are performed directly by the publisher (contact@affinix.bio).

Affinix is hosted on the serverless edge execution infrastructure of Cloudflare, Inc. (101 Townsend St, San Francisco, CA 94107, USA), through its French subsidiary Cloudflare France SAS (6 Place de la Madeleine, 75008 Paris, France; SIREN 883 044 646). For all European Union and European Economic Area (EEA) users and sessions, edge worker computation and data transit are exclusively routed through European Union data centers.

When you execute a simulation, your selected regimen, patient parameters, and organ function indicators are transmitted via encrypted HTTPS directly to the serverless Worker for mathematical integration. These parameters are held in transient memory solely for numerical solving. To optimize execution and prevent repeated computations, mathematical trajectories may be ephemerally cached in volatile Cloudflare Worker isolate RAM and edge cache for up to 24 hours, keyed strictly by an irreversible SHA-256 hash of anonymous numerical parameters, with zero persistent disk logging, zero database storage, and zero patient identification. For authenticated users, basic account identity, professional credentials, role assignments, and capability entitlements are maintained in a secure, encrypted European Union Cloudflare D1 database governed by European GDPR standards.

Affinix uses the client browser local storage mechanism under the designated key "affinix-sim-storage-v3" solely to allow the user to retain their configured scenario across browser refreshes on their own physical device. This data is strictly local, never synchronized to any Affinix server, and may be purged by the user at any time through their browser settings or the platform reset option.

Affinix deploys zero tracking cookies, advertising trackers, third-party analytics scripts, or commercial profiling identifiers. For authenticated sessions, strictly necessary first-party cryptographic session cookies (configured as HttpOnly, SameSite=Lax, and Secure) are issued solely to maintain session integrity, prevent cross-site request forgery, and verify access entitlements. Technical performance, network latency, and Core Web Vitals are monitored using Cloudflare privacy-first, cookieless edge Real User Monitoring (RUM) beacon with zero advertising tracking, behavioral profiling, or cross-site identifiers.

To protect authentication endpoints and email dispatchers from automated credential stuffing, distributed denial of service, and brute-force attacks, Affinix utilizes Cloudflare Turnstile bot detection. This evaluation operates on privacy-preserving cryptographic challenge tokens without persistent tracking, device fingerprinting, or user browsing history profiling.

When requesting passwordless sign-in or account verification, a temporary six-digit authentication code is dispatched to the user's institutional or designated email address via Cloudflare Email Service. Email addresses are processed solely for the transmission of transactional security passcodes, expire automatically after five minutes, and are never enrolled in commercial mailing lists or disclosed to third parties.

The Affinix infrastructure is deliberately designed not to process or host identifiable personal health data and does not claim certification as an accredited French health data host (hébergeur de données de santé) pursuant to Article L1111-8 of the French Public Health Code. Storing real patient files or clinical notes on this platform is unauthorized and unlawful.

All patient profiles configured in Affinix (weight, age, eGFR, organ clearance indices, genotypes) must represent synthetic archetypes, textbook physiological cohorts, or fully de-identified experimental cohorts that cannot be correlated with any identifiable natural person.

Any user who inputs identifiable clinical records or personal health information into scenario names or shareable URL hashes assumes sole civil, regulatory, and criminal responsibility under the GDPR and applicable health privacy statutes, and shall fully indemnify and hold harmless the publisher against any third-party claims, regulatory fines, or administrative proceedings.