X-RAIS
X-RAIS is a medical image analysis AI platform from Laife Reply that reads mammograms, chest X-rays and brain MRI scans, flags suspicious findings and drafts reports for radiologists inside existing PACS and RIS workflows.
What is X-RAIS?
X-RAIS is a medical image analysis AI platform based on neural networks and radiomics, developed by Laife Reply, the healthcare unit of the Italian group Reply. Rather than one general-purpose model, it is verticalised: specialised by diagnostic method, such as radiography, ultrasound, CT and MRI, and by specific anatomical region. Its declared role is that of a cognitive assistant supporting physicians at the diagnosis and reporting stages, with the stated aim of reducing the number of misdiagnoses and improving the efficiency of the whole diagnostic process. It is not presented as an autonomous decision system.
Three specialisations are described as delivered. Mammography came first, built with the scientific partners IRCCS IEO, ICS Maugeri and ULLS2 Marca Trevigiana, with dedicated models for suspicious microcalcifications, radio-opacities and suspicious breast lesions; the page notes that breast cancer is the most frequent cancer in women, at 30.3% according to AIRC. The second is chest radiography, developed inside the ALFABETO project on Covid-19 with ICS Maugeri and the University of Pavia, aimed at triaging symptomatic patients at home, combining the image with instrumental and anamnestic data to grade severity. The third is 3D brain MRI, where the platform segments glioma and classifies its severity. Histopathology imaging and the identification of new biomarkers are described as research and development, on a path towards precision medicine.
Architecturally, the platform is a set of technology modules, APIs and Web Services, that connect to diagnostic systems such as RIS and PACS through DICOM node integration, alongside anonymisation and sensitive-data protection routines and vertical modules for value-based healthcare. Images acquired by diagnostic equipment are analysed and the identified lesions and reports are sent back to the PACS, while examinations returning a suspicious result are automatically escalated in priority for specialist staff. A July 2025 announcement adds a multi-agent system: one agent helps generate the report enriched with the patient's history, a second assigns scores to prioritise the most critical cases, and integration with existing healthcare systems is described through APIs compliant with FHIR standards.
Cognitive and AI services are delivered through cloud architectures, with on-prem installation or use possible in specific cases. The design is explicitly human-in-the-loop: the radiologist stays at the centre of the process. The product page itself was published in December 2023.
What it does
- Automatically analyse medical images acquired by diagnostic equipment
- Flag and classify suspicious areas: microcalcifications, radio-opacities and suspicious breast lesions
- Segment glioma on 3D brain MRI and classify its severity
- Grade the severity of a condition from a chest X-ray combined with clinical data
- Return identified lesions and reports to the PACS
- Automatically raise the priority of examinations with a suspicious result, for smart triage
- Assist report writing by drawing on the patient's history
When to use X-RAIS / When not to
A quick filter to help you decide if X-RAIS is the right fit.
When to use X-RAIS
- High-volume radiology and diagnostic imaging departments where the number of examinations outpaces the available reading time
- Organised mammography screening programmes looking for support on microcalcifications, radio-opacities and suspicious breast lesions
- Hospitals already running a PACS and a RIS, with the IT capability to host an additional DICOM node
- IRCCS research hospitals, public and private hospitals and health authorities, the customer profile Laife Reply describes for its healthcare unit
- Clinical research teams working with scientific partners, in the mould of IRCCS IEO, ICS Maugeri, ULLS2 Marca Trevigiana and the University of Pavia
When not to use X-RAIS
- Patients and members of the public: X-RAIS is built for medical staff and has no consumer-facing access
- Anyone hoping to sign up online, since there is no self-service account, no trial and no self-deployment; every engagement goes through Reply
- Teams needing modalities beyond the three delivered specialisations, as histopathology imaging is still described as research and development
- Departments looking to automate reporting without a clinician in the loop: the human-in-the-loop design deliberately keeps the radiologist responsible
- Buyers who must budget from a published price list, because no amount is public and every engagement has to be quoted
How to use X-RAIS
A typical end-to-end flow, from setup to results.
- Check the fit first: X-RAIS is sold to institutions, and the delivered specialisations are mammography, chest radiography and 3D brain MRI
- Make contact through the Laife Reply contact form or the Reply contacts page, since there is no online sign-up and no self-service account
- Scope the project with Laife Reply, including which vertical modules apply and whether cloud or on-prem use is appropriate
- Agree commercial terms directly with Reply: usage is offered on a pay-per-use or periodic subscription basis, with nothing quoted publicly
- Plan the integration with hospital IT and medical physics, as X-RAIS connects to the RIS and PACS through DICOM node integration
- Connect the technology modules, APIs and Web Services, and where relevant the FHIR-compliant APIs to existing healthcare systems
- Route the images acquired by diagnostic equipment to the platform for automatic analysis
- Collect the identified lesions and generated reports back in the PACS, where radiologists read them in their usual environment
- Rely on smart triage: examinations with a suspicious result are automatically promoted to high priority for specialist staff
- Keep the radiologist in the loop, with the reporting and scoring agents acting as support rather than as the decision-maker
Pros & Cons
Pros
- Backed by a listed European group: Reply S.p.A. trades on the Euronext STAR Milan segment (ISIN IT0005282865, ticker REY) and reports more than 17,200 people across 16 countries
- Clinical grounding through named partnerships with IRCCS IEO, ICS Maugeri, ULLS2 Marca Trevigiana and the University of Pavia
- Three distinct specialisations already delivered, across three different imaging modalities
- Designed to sit inside the existing workflow through PACS, RIS, DICOM and FHIR, rather than as an isolated tool
- Explicit human-in-the-loop approach that keeps the radiologist as the decision-maker
- On-prem installation or use possible in specific cases, for institutions that cannot move data outside their own walls
- European publisher with a controller established in Italy and designated DPOs, plus a first-party scientific publication on microcalcification annotation available as a PDF
Cons
- No public pricing of any kind: no amount, no price grid and no pricing page, checked across the 1,395 URLs of the English sitemap and on both the English and Italian product pages
- No free trial and no openly accessible demonstration; the only way in is a contact form
- No API documentation is published, even though the platform advertises API and Web Services modules and FHIR-compliant APIs
- No privacy policy and no data processing agreement specific to the product: the site's legal documents cover the website and the client relationship, not the processing of clinical data
- No hosting country or region is disclosed and no subprocessor list is published
- No certification such as ISO or SOC 2, no CE marking and no medical device status is mentioned on the product page, and no interface language is declared
- The contracting entity is never named, the Privacy Notice (Client) referring only to "the Reply Group's Company, holder of the contractual relationship", and the product page has not been dated since December 2023, the most recent item being a July 2025 announcement
Pricing & Plans
No price is published for X-RAIS anywhere on reply.com: neither the English nor the Italian product page carries an amount, and the site has no pricing page. The publisher states only that the platform "can be used on a pay-per-use or periodic subscription basis", without designating either arrangement as the standard one, and adds that in certain cases the architecture also allows on-prem installation or use. There is no free plan and no free trial. A budget can therefore only be established through a quotation obtained directly from Reply.
Data, GDPR & hosting
A consolidated view of how X-RAIS handles your data.
GDPR overview
GDPR compliance is asserted at publisher and website level, never in a document specific to X-RAIS. The product page carries a single sentence: X-RAIS "guarantees data integrity and confidentiality, which are crucial in light of the new privacy regulations in force (GDPR)". Behind it sits the group's apparatus: a controller established in Italy, so no Article 27 representative is required, and three designated DPOs, dpo.it@reply.it for Italy, dpo.de@reply.de for Germany and Poland, dpo.uk@reply.com for the United Kingdom. The Privacy Notice (Client), issued pursuant to Section 13 of the EU and UK GDPR and last updated in November 2022, lists access, erasure, rectification, integration, restriction, objection and complaint to a supervisory authority; rights are exercised at client.privacy@reply.com, and at marketing.privacy@reply.com for the site, whose Privacy and Cookie Policy dates from December 2025. No product-specific DPA is published.
Who owns the data?
No privacy policy is published for X-RAIS itself, so no document sets out who owns or controls the clinical images an institution would process. Two site-level documents exist. The Reply Website/Platform Privacy and Cookie Policy, last updated in December 2025, names Reply S.p.A., registered office Corso Francia 110, Turin, as data controller, and covers the reply.com website only. The Privacy Notice (Client), last updated in November 2022, designates as controller "the Reply Group's Company, holder of the contractual relationship" with the client, without ever naming that entity, and concerns client contact data rather than patient imaging. On the product side, the only published statement is that the architecture includes routines for anonymisation and protection of sensitive data.
Reuse rights
The website policy describes technical, profiling and techno-functional cookies plus ZoomInfo company identification, with prior consent required for profiling. The Privacy Notice (Client) covers contact data of the client's employees and contractors, processed to perform the contract, meet legal obligations and pursue the group's legitimate interest. Neither document says anything about what happens to the clinical images a hospital would submit through X-RAIS. The models themselves were trained on anonymised clinical data supplied by named scientific partners, not by customers: Reply states that several scientific partners provided "both the anonymized clinical data and the know-how for its annotation". No commitment is published on whether customer data is or is not used for training, and no opt-out mechanism is documented. Reuse rights over analysis outputs are not addressed anywhere in public, and there are no terms and conditions on the site to settle the question.
Data retention & training
Hosting summary
No hosting country and no hosting region is announced for X-RAIS. The only published facts are that cognitive and artificial intelligence services are offered through cloud architectures and that, in certain cases, the specific architecture of the platform also allows on-prem installation or use, which means clinical data can in principle remain inside the institution, subject to project scoping. No cloud provider is named, no data centre location is given, and no subprocessor list is available. The only jurisdiction that can be established is that of the publisher rather than of the data: Reply is established in Italy, and the controller named for the reply.com website is Reply S.p.A., with a registered office in Turin. One point deserves care. The reply.com website itself is served from an AWS IP address located in Ireland (99.80.3.188), but that is an infrastructure fact about the corporate website and says nothing about where clinical images processed by X-RAIS would be stored. For a hospital, hosting location and data residency have to be obtained contractually from the publisher; they cannot be read off the public pages.
Things to keep in mind
Risks and trade-offs to weigh before adopting X-RAIS.
- The budget cannot be estimated before a quotation: no amount is published, and both announced arrangements, pay-per-use and periodic subscription, are left unpriced
- The contracting legal entity is never named; the Privacy Notice (Client) points only to "the Reply Group's Company, holder of the contractual relationship", which must be pinned down before signature
- No CE marking, medical device status or security certification such as ISO 27001 or SOC 2 is mentioned for X-RAIS, a decisive point for diagnostic support software in Europe
- No privacy policy or data processing agreement covers the product itself, leaving the processing of clinical images contractually undefined by the public documents
- No hosting country or region is announced, even though health data is at stake, and no subprocessor list is available for a vendor security assessment
- APIs are advertised but no documentation is published, so the true integration effort cannot be assessed before the project starts
- Automation bias is the human risk to manage: smart triage scores and generated reports are only safe if readers keep questioning the flags instead of deferring to them, and since the product page dates from December 2023, the current state of the specialisations should be confirmed with the publisher
Setup & Integrations
Technical difficulty
This is a system integration, not a software installation. X-RAIS connects to the institution's PACS and RIS through DICOM node integration, exposes API and Web Services modules, and offers FHIR-compliant APIs for connection to existing healthcare systems; on-prem installation or use is possible in specific cases. There is no self-service path: deployment runs as a project with Laife Reply, and no public technical documentation exists to prepare the work beforehand. Expect to involve hospital IT, the PACS and RIS administrators and information security, and to plan the integration jointly with the publisher.
Deployment
Behind X-RAIS
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does X-RAIS do?
Which imaging modalities does it cover?
Does X-RAIS replace the radiologist?
How does it integrate with hospital systems?
How much does X-RAIS cost?
Can it be installed on our own premises?
Is there a free trial or a demo?
Who publishes X-RAIS?
Is there a documented API?
Where is the data hosted?
Should you pick X-RAIS?
X-RAIS is clinically substantial and commercially opaque, and a buyer should weigh those two things separately. On the clinical side the substance is real: three delivered specialisations across three imaging modalities, named scientific partners in IRCCS IEO, ICS Maugeri, ULLS2 Marca Trevigiana and the University of Pavia, a first-party scientific publication on microcalcification annotation, and a design that keeps the radiologist in the loop rather than automating the decision. The architecture is built for hospitals that already run a PACS and a RIS: a DICOM node, API and Web Services modules, FHIR-compliant APIs, and results pushed back into the PACS where radiologists already work.
On the commercial side, almost nothing is public. There is no price, no free trial, no demonstration, no API documentation, and no privacy policy or data processing agreement specific to the product. The legal documents on reply.com cover the website and the client contractual relationship, not the handling of clinical images. Nor is there any statement of where data is hosted, any subprocessor list, any certification, or any mention of CE marking or medical device status.
The realistic buyer is an institution rather than an individual: a radiology department or health authority with the IT capability to integrate a DICOM node and the procurement capability to run a scoped project with Laife Reply. Before signing, three questions deserve to be put in writing: which legal entity actually contracts, since the Privacy Notice (Client) names only "the Reply Group's Company, holder of the contractual relationship"; what the regulatory status of the software is for diagnostic support in Europe; and where clinical data is processed and stored. The product page dates from December 2023, so it is also worth asking the publisher which specialisations are current today.
- Choosing a selection results in a full page refresh.
- Opens in a new window.