Frequently asked questions

Synaris is a workspace for organising case information, exploring clinical hypotheses that need further assessment, and preparing documentation. It brings together clinical reasoning support, notes, case records and appointments. During the Beta, all of these activities must use entirely fictional cases.

It is designed for psychologists, psychotherapists and mental-health professionals. The private Beta invites professional colleagues to assess its clarity and usefulness in simulated work. It is not a public self-diagnosis service or a patient-facing assessment tool.

Synaris helps keep the case narrative, clinical indicators, unanswered questions, hypotheses, notes and appointments connected. Its aim is to reduce fragmented working and make the next questions easier to identify. It does not claim demonstrated clinical benefits or proven time savings.

It provides a structured workflow: reviewing indicators, recording clinical context, considering hypotheses, reviewing the output, documenting the case and arranging follow-up. AI features have defined tasks within that workspace; they are not the entire product. The structure does not remove the possibility of AI errors.

No. Its hypotheses are suggestions for assessment, not established diagnoses. Interpretation, diagnostic and treatment decisions, and responsibility for documentation remain with the clinician. Completing a screen or obtaining agreement between AI outputs is not clinical validation.

Start with an entirely fictional case. Enter the narrative, review the proposed indicators, and record duration, functional impact and relevant context before examining the hypotheses. You can then use the Reviewer and Explorer, consult SynPedia, transfer a draft to a clinical note, save and reopen it, and create a fictional follow-up appointment.

Check that each proposed indicator is supported by the narrative. Correct negations, remove irrelevant items and add missing information manually. An unselected item does not necessarily mean it has been assessed and ruled out.

They represent different states of knowledge. An absent finding has been explicitly denied; a finding that has not been assessed is unknown. Duration, impairment and context fields help preserve those distinctions. Do not enter negative answers simply to complete the screen.

Earlier outputs may be marked as needing an update. Recalculate the hypotheses and refresh the Reviewer and Explorer if you are using them. An earlier response retained for comparison does not automatically reflect the revised case; check it before transferring it to a note.

It provides shortcuts to the main areas, operational information and appointments. Use it to resume a case, open a note or check the diary. What you can see depends on your role, the cases you can access and the information already saved.

In the current workflow, AI suggests hypotheses to consider from information the clinician has entered and checked. It does not independently observe the consultation and may lack details, context or nuances the clinician has not supplied. The clinician must assess, correct or discard each suggestion.

Evidence refers to case information you have supplied and checked, not observations made independently by Synaris. References to criteria may help the clinician assess the case; they do not certify that diagnostic criteria have been established. An AI-generated question or missing-information item is not automatically an unmet DSM criterion.

They are alternative explanations to consider against the case. Synaris may present supporting points, limitations and questions that help distinguish them. An item appearing in the list does not establish a diagnosis or comorbidity, and an omitted condition has not necessarily been ruled out.

Synaris may highlight points to clarify and suggest follow-up questions. A gap is neither an automatically unmet criterion nor an “almost established” diagnosis. The clinician decides what to ask and how to interpret the answer in context.

Read them against the narrative and any explicit denials. An alert alone does not establish risk, and the absence of an alert does not mean that a complete risk assessment has taken place. Synaris is not an emergency service and does not replace appropriate clinical care pathways.

Available views may show supporting information, contradictions, gaps and follow-up questions. These help you examine and challenge an output. An AI-generated explanation may itself be wrong: it is neither proof that the hypothesis is accurate nor a certified account of the model's internal reasoning.

Synaris does not claim complete or uniform DSM coverage. The catalogue, information cards and suggested hypotheses have different purposes and limits. Documented technical checks and fictional test cases are not prospective clinical validation. This Beta is not presented as evidence of diagnostic accuracy.

SynPedia provides draft information cards associated with a selected hypothesis when a matching entry is available. The Catalogue lets you consult profiles and indicators. Neither replaces official manuals or professional training; check the cited sources. A missing card neither confirms nor excludes a condition.

The current workflow uses AI to extract indicators, generate clinical hypotheses, and provide Reviewer and Explorer contributions. There are also writing tools and optional assistance with note sections such as mental state examination and 5P formulation. Opening “Clinical hypotheses” may itself trigger an AI request when a new response is needed.

You can write notes manually, manage accessible cases and use the diary without these contributions. The Reviewer, Explorer and writing suggestions are optional. The current “Clinical hypotheses” step does require AI.

The Reviewer critically examines the current hypothesis for consistency, contradictions and questions that need clarification. The Explorer broadens the discussion with alternatives and further questions. Neither confirms a diagnosis. The Explorer in the ordinary workflow is separate from administrator-only experimental tools.

No. It brings together AI contributions already generated in the workflow. It does not record a human consultation or provide an independent validation. You can select useful sections, review them and prepare a note draft. The “Clinical Team” laboratory is a separate experimental area.

Yes. It may misread a denial, ask about something already explained, or introduce an unsupported interpretation. It works only with the information supplied: it does not directly observe the consultation or clinical relationship. The clinician can add context and nuances absent from the system, and must check, correct or discard unsupported content. Report errors using fictional examples only.

Requests require an active practice AI policy and the relevant permissions. Privacy checks, usage limits or technical issues may also block a request. Accepting the personal Beta notice and activating the practice policy are separate steps; neither permits real patient data. If a problem persists, record the message and contact the Beta coordinator.

When you use an AI feature, the backend sends the material needed for that task to the configured provider; the existing integration uses OpenAI. The prepared version includes filters intended to reduce identifiers in the outgoing copy, but they do not guarantee complete anonymisation. Those filters alone do not establish zero retention, the absence of international transfers or suitability for real patient data.

You can find accessible cases, open their details and review linked notes and appointments. Notes support initial assessments and follow-up sessions, with explicit saving, reopening and editing. During the Beta, all names, codes and case content must relate to fictional people.

Yes, through the available explicit actions and by selecting relevant material. The transfer creates a draft: check its contents and the linked case, then complete the save. It does not turn hypotheses into established diagnoses or save the note on its own.

No. A note draft is temporarily held in the browser session and may be lost when it closes. The diagnostic workflow can also retain a draft on the server. To find a note in the records later, explicitly save it, check the result and reopen it.

Writer is a supplementary tool for preparing and editing professional drafts from the available material. It supports voluntary transfer to a note, RTF export and printing. Always review the text. In the prepared version, if refreshing the page loses the protected transfer context, reopen Writer from the note; the block prevents generation without that context.

There are structured clinical notes, optional sections such as MSE and 5P, and references to selected screening instruments in supplementary areas. This is not a complete suite of tests that can be administered and scored automatically. Beta feedback questionnaires serve a different purpose: collecting views on the application.

It manages appointments, filters and links to cases. You can prepare a Google Calendar event or download an ICS file. These are copies or imports, not continuous two-way synchronisation. Check dates and times, and use fictional content in external calendars too.

Specific outputs include printing a selected summary, RTF export and printing in Writer, appointment ICS files, and a restricted administrative JSON export. Your browser may let you save printed output as PDF. Synaris does not promise universal export from every screen or compatibility with every clinical records system.

No. These are supplementary tools with different availability and persistence. The case map uses a session draft, not a saved clinical record. Use the clinical note to document the work and check anything you transfer to it.

Clinicians use the clinical functions permitted to them. A Practice Administrator also manages practice settings, users and policies. Technical platform administration is separate. An administrative role does not automatically grant access to every case.

A case can be assigned to one or more professionals in the same practice. In the prepared, locally verified version, clinical access requires both practice membership and a current assignment. Revoking it prevents subsequent access to that case; it does not remove copies already exported. The controls also apply to AI requests linked to the case.

In the locally verified implementation, an administrator must also be assigned to a case to read its clinical content. Assignments can be managed through a separate administrative register. Administrative export follows the relevant access scope rather than providing unrestricted clinical access.

The external Beta is planned around a separate practice for each tester. The application also supports inviting colleagues into an existing practice, but that is a different workflow. Follow the instructions provided for your account and do not use another tester's credentials or personal access links.

No. This Beta permits entirely fictional cases only. Removing identifiers, using a code or obtaining consent does not change that scope. Do not enter real clinical narratives or notes even when the patient's name is absent.

Role, practice and case-assignment checks have been developed, together with measures intended to reduce identifying information sent to AI features. Logging and error handling have also been strengthened and checked locally. This does not establish that every change is active in the public environment or that the system is ready for real health data.

No. Encryption of the title and content of newly created clinical notes has been developed and checked locally. Identity data, other clinical fields and older notes are not all covered. The full protection required for an environment handling real health data is not ready. Use fictional cases only in this Beta.

No. A production key-management service and the other protections needed for real health data are not yet complete and verified. Choosing an infrastructure provider alone would not make this Beta suitable for real patient cases. Use fictional cases only.

It clears the relevant temporary material in your browser. It is not the same as deleting saved notes, server-side drafts, your account or backups. Save any work you want to retain first. Other deletion requests require the dedicated controls and, where necessary, help from the coordinator.

It visually covers the workspace, which can help when someone else can see your screen. It does not sign you out, save your work or replace device security. Sign out when you finish testing.

No such claim is being made. Technical measures and draft notices exist, but production settings, contracts, retention and the conditions for handling real health data still require further review. Even a fictional-case Beta processes professionals' personal information, including email addresses and account details.

It evaluates the clarity, operation and perceived usefulness of the workflow using fictional cases, not diagnostic validity. Participation is limited to invited colleagues. Use the link and email address agreed with the coordinator; this is not unrestricted public registration. The application includes sign-in, email verification and password recovery workflows.

Participation in the private Beta is free of charge. The Beta period and length of access will be communicated to participants and may be adjusted during testing. Free Beta participation does not imply that the future commercial service will be free.

Try one or two fictional cases, review the workflow and describe what helps or causes confusion. Useful feedback identifies the page, action, expected outcome and observed result, together with browser and device details. You do not need to obtain a predetermined diagnosis or adjust the case to make the system produce one.

The Beta is evolving: features and interfaces may change, and they are not guaranteed to remain identical. Do not use it as your only record store. End-of-Beta arrangements, any retention or deletion of fictional content, and what happens to accounts have not yet been decided; they will be communicated before participation. No automatic move to a paid plan is promised.

Synaris runs in a web browser; no native app is promised. The interface includes layouts for smaller screens, but comprehensive compatibility testing across all browsers, tablets and phones is not documented. Use the environment specified by the coordinator, and report your device, browser and screen size if something does not work as expected. Offline operation is not promised.

The interface provides English and Italian, with light, dark and system theme options. Changing language does not automatically translate existing notes or narratives, or guarantee identical catalogue coverage in both languages. Reduced-motion preferences are also available.

Commercial plans will be announced before the public launch. Monthly, quarterly or annual prices, discounts and renewal terms are not stated until approved. Taking part in the Beta is not described as a commercial subscription.

For enquiries and issues, contact adminsynaris@gmail.com. The app includes links to Beta feedback questionnaires in English and Italian. Do not send passwords, tokens or real patient material. Response times and emergency support are not promised. Operating guides will be made available only after they have been updated and approved.