A platform built around the real structure of the therapy process,
not around a generic note.
It brings order to thinking about the case, not only to the paperwork. The time that used to go on administration comes back to therapy, to the things there was never room for.
Someone among your fellow therapists thought Therapy Support is an application for you.
Therapy Support is not a transcription tool or an AI notepad.
The platform divides the work into the elements known from clinical practice: the initial consultation and the demographic and developmental history, problems and therapy goals, clinical situations, automatic thoughts and cognitive distortions, ABC analysis, case conceptualisation with the vicious cycle model, and in work within schema therapy also schemas, modes with the mode map, coping styles and unmet emotional needs.
Every session is mapped onto the structure you know from practice. The features reflect the therapeutic process rather than redefining it.
The application reads the transcript for the structure of the session: it marks distortions, behaviours and the techniques used. The same annotations then feed the note and the conceptualisation.
The view comes from the application. The patient data is fictional.
The work happens between sessions rather than during them.
The session transcript becomes a structured note following the structure of a CBT session: agenda, homework reviewed, interventions detected, new assignment, summary. You check it, correct it, save it. Fifteen to twenty minutes of work becomes two to three minutes of editing.
The catastrophising thread from sessions 3 and 7. Schemas, modes and coping styles tracked from session to session. Your patient fills in the GAD-7 on Sunday, on Monday you see the drop from 14 to 10, and you begin the first minute of the session differently.
A cognitive diagram, vicious cycle model and ABC analysis, and in schema work the mode map and unmet emotional needs - generated from session content and updated with each new session. In session 6 the patient mentions his father for the first time, you add a hypothesis, and it appears alongside the existing elements of the model.
You open the patient record and see the homework from the previous session, the questionnaire score from this morning, the thread from session 3. Before supervision: an extract of key threads, hypotheses and moments of change from selected sessions, in a single document.
The platform organises material from the whole therapeutic process and prepares proposals. Everything produced in the application is a proposal for review, not a finished determination.
Data from therapy sessions is a special category of data. Compliance and privacy are part of the design, not an addition to it.
Always Encrypted at database column level, with keys held in Azure Key Vault.
Names, national ID numbers, phone numbers and emails are detected and removed before every model call.
Our own Azure infrastructure in inference mode, not the public ChatGPT or Claude API.
A clause in the contract with the provider, not a marketing claim.
Azure: Poland Central, Sweden Central, West Europe. Compliant with the GDPR and the AI Act.
Full history export, permanent deletion on request, a complete audit log, and 2FA at sign-in.
A study of the platform’s effectiveness carried out within the SMART NCBR consortium. It covers 360 synthetic clinical cases, with randomised participants and blind assessment of results.
In the report “AI in Psychotherapy” (module 4, a map of psychotherapy tools) Therapy Support was recognised as a promising solution on the Polish market.
Partners: Microsoft for Startups · Polish Academy of Sciences · SWPS Booster · Lublin Medicine · Sanitas Medical Centre · Poznań Science and Technology Park
The simplest route is an onboarding call: we show the application on an example close to your own practice and answer questions about data security.