Software for CBT therapists: what it must do and how to choose

CBT therapist software is a tool that, beyond holding the clinical record, keeps track of what gives cognitive behavioural therapy its structure: the case formulation, measurable goals, repeated measures, homework between sessions, exercises (thought records, exposure hierarchies) and the end-of-therapy summary. Standard practice software mostly handles the diary and invoices. In CBT the real question is different: can you see in two minutes where each client stands on each goal?

What CBT asks of a tool

General software stores notes. CBT needs those notes to connect from one session to the next. Here are seven things you should find without digging.

• The case formulation. The working hypothesis (maintaining factors, beliefs, safety behaviours) should stay visible and dated, because it changes over time. The full template is in our article on the 5 Ps case formulation.
• Goals. Agreed with the client, concrete, with a status (met, ongoing, dropped, reworded) and the date of the last review.
• Repeated measures. The same indicator taken the same way each session or at set intervals: a 0 to 10 anxiety rating, panic attacks per week, a questionnaire score if you are allowed to use it. The tool should keep the history, not just the latest value.
• Homework. What was agreed, what was done, what got in the way. Homework not reviewed at the next session loses its effect, so the tool should remind you.
• Thought records. Situation, automatic thought, emotion and intensity, alternative thought. For follow-up, what matters is not each sheet but the recurring thoughts and how belief ratings change.
• Exposure hierarchy. Situations ranked by expected anxiety, the step reached, and anxiety at the start, peak and end of each exercise. How to write an exposure session note is covered in our CBT session note template.
• Ending therapy. What changed, goals met, early warning signs of relapse and the client's plan. See our end-of-therapy summary template.

What each type of tool does

The information below comes from the vendors' websites, read on 5 November 2026. We have not tested these tools. Anything not shown on a vendor's site is marked "not stated on its site", which does not mean the feature is missing.

• General practice software (example: Doctolib). Strength: booking and online appointments, which its French site highlights for psychologists ("Agenda, Rendez-vous & Développement de Patientèle"). CBT follow-up (goals, measures, exposure): not stated on its site.
• CBT-oriented practice software (examples: Psy.link Pro, PsyLib, both French). According to its site, Psy.link Pro combines booking, invoicing, video sessions, an exercise library and questionnaires sent to clients and scored automatically, for €49 a month, with data hosted in France. Thought records, exposure hierarchies, goals and clinical notes: not stated on its site. According to its CBT guide, PsyLib offers a customisable CBT note template, PHQ-9 and GAD-7 tracking on a chart and a client space for homework, with HDS-certified hosting. Shared strength: everything sits next to the diary.
• Between-session platform (example: Intervall, French). According to its site, it lets you assign clinical measures (PHQ-9, GAD-7) and CBT exercises and follow clients between sessions, and it describes itself as GDPR-compliant. Health data hosting, session notes and progress summary: not stated on its site.
• Client self-monitoring app (example: Feel, French). According to its site, Feel is a programme of CBT exercises to help people out of depression, helps users track their emotions, and is "not a medical device". Its strength is day-to-day client engagement. Sharing data with the therapist: not stated on its site, so you only see what the client shows you.
• AI note assistant (examples: Heidi, Upheal). According to its UK site, Heidi turns a session into structured progress notes, risk assessments and letters in your preferred modality. According to its UK page (the company is based in the US), Upheal automatically generates progress notes and, on its paid plan, offers treatment plans with SMART goals. Repeated measures and exposure hierarchies: not stated on their sites. Hosting location: not stated on the pages we read for either, so ask. Our comparison of AI note-taking software covers the options in more detail.
• Spreadsheet. Strength: flexible, free, ideal for an anxiety curve or a hierarchy. Limits: no link to your notes, re-entry every session, and a file of health data on a laptop or consumer cloud raises security questions.

In practice, many CBT therapists combine two tools: practice software for booking and invoices (see our guide to psychologist practice software), and a tool that holds the clinical memory of each client.

Three rules to check before adding clients

Questionnaires are copyrighted

A published questionnaire is not free to copy by default. For the BDI-II, Pearson states that "Any other use of the BDI-II via telepractice requires prior permission from Pearson" beyond its own platform (Pearson Clinical UK). Copying the items of a protected questionnaire into an online form, spreadsheet or app may therefore fall outside the licence. We advise checking each questionnaire's licence on its publisher's site, using the authorised tool where one exists, and storing only the date, the questionnaire name and the total score in your software.

Data protection: health data and processors

In the UK, the ICO explains that the UK GDPR singles out some personal data as more sensitive "and gives them extra protection", and health data is one of these special categories (ICO). In Ireland, the DPC states that processing these special categories "is prohibited, except in limited circumstances set out in Article 9 of the GDPR" (DPC). In France, the CNIL tells health professionals in private practice that a provider processing data for them must, as a processor, guarantee a level of security suited to the risk, and that clients must be informed (CNIL, in French). In practice: ask for the data processing agreement, the hosting location and, for France, whether the host is HDS-certified (a French framework, not a legal requirement elsewhere).

The French code of ethics for psychologists, article 23

For psychologists practising in France, article 23 of the 2021 code of ethics states (our translation) that the psychologist collects, processes, files and archives their personal notes and related data so as to protect people's privacy and respect professional confidentiality, and that any data used for publication or communication must be fully anonymised (Code de déontologie des psychologues, consolidated 9 September 2021). Our advice: before sending worksheets to a third-party app or exporting data to a shared spreadsheet, know where your notes go and who can see them.

Selection checklist for CBT therapists

• Memory across sessions: before a session, does the tool show goals, latest measures and current homework without opening five notes?
• Modality: does the note read like CBT (situation, thought, emotion, behaviour, homework) or like a generic summary?
• Measures: can you track your own ratings (0 to 10, frequency) and see their history? Are the questionnaires on offer properly licensed?
• Exposure: can you find the hierarchy and the step reached from one session to the next?
• Ending: can you produce an end-of-therapy summary or letter from the whole course of therapy?
• Data: hosting location, data processing agreement, what happens to audio if the tool transcribes, whether data trains a model.
• Client side: if you want a client app, what do you see of what the client fills in, and who hosts it?
• Exit: full export of your records if you switch tools.

The test to run during the trial

A demo always shows the ideal case. We suggest this one-hour test with a fictional client.

• Step 1. Create a fictional client with a short formulation and three goals.
• Step 2. Type, dictate or role-play three sessions: an assessment, a cognitive restructuring session with a thought record, and an exposure session with anxiety ratings.
• Step 3. Before session four, ask the tool where the client stands. Do you find the goals, the rating trend and the agreed homework?
• Step 4. Change one goal (dropped or reworded) and check that the summary follows.
• Step 5. Ask for an end-of-therapy summary and an export. If you have to rewrite everything, the tool does not keep the memory of the therapy.

Fictional example: a session note and the expected summary

Fictional client: Lea R., 29, social anxiety, struggles to speak up in meetings. All values are invented and come from subjective ratings agreed with the client, not from a questionnaire.

Session 6 note

• Homework review: two of three thought records completed. Recurring thought: "They'll see I don't belong here." Belief rated 70 out of 100 before restructuring, 45 after.
• Weekly measure: meetings avoided, 2 out of 5 (4 out of 5 at session 1).
• In session: imaginal exposure at step 4 of the hierarchy (asking a question in a team meeting). Anxiety 65 at start, 80 at peak, 35 at end. Safety behaviour noted: rereading her question three times before speaking.
• Homework agreed: ask a question in Thursday's meeting without rereading it first, rate anxiety before and after.

Progress summary the tool should give after session 6

• Formulation (revised at session 4): core belief "I don't belong here", maintained by avoiding meetings and over-checking.
• Goal 1, speak at least once per meeting: ongoing, step 4 of 7 on the hierarchy.
• Goal 2, reduce meeting avoidance: ongoing, from 4 out of 5 (session 1) to 2 out of 5 (session 6).
• Goal 3, return to sport: postponed at the client's request at session 3.
• Belief worked on: from 85 out of 100 (session 2) to 45 out of 100 after restructuring (session 6).
• Current homework: question in a meeting without rereading, to review at session 7.
• Topics for next session: homework review, move to step 5 if end-of-exercise anxiety stays under 40, work on rereading as a safety behaviour.

This second block is what sets a CBT-ready tool apart: it is built from every session, not just the last one.

Common mistakes

• Choosing on the booking demo. A diary is easy to replace; the clinical memory of a course of therapy is not.
• Copying a protected questionnaire into a homemade form. Check the licence first.
• Stacking client apps. Each app adds a host and another place where your data slips out of view.
• Logging measures without a scale. "Anxiety 6" means nothing unless you know whether it is out of 10 or 100.
• Leaving goals frozen. A dropped or reworded goal must show as such, or the end-of-therapy summary will be wrong.

Why Delta fits CBT follow-up

Delta is the clinical assistant for mental health professionals: it keeps the memory of each client. During the session it transcribes what is said, or you dictate your observations afterwards. It generates the session report (short summary, detailed summary, topics for next session) taking your profession and approach into account, CBT included, and updates the client record (progress summary, goals tracked). You find the whole course of therapy and how it evolved in one place. Data stays in Europe, in France, on infrastructure certified for health data; the client's name is pseudonymised before the AI and no audio file is kept.

Delta does not score questionnaires and handles neither booking nor invoicing: it works alongside your practice software. Learn more in what Delta is, on the Delta for psychologists page and in our clinical document templates. The trial lasts 14 days with no card required; after that, the Free plan covers 10 sessions a month and the Pro plan costs €35 a month including VAT with no commitment (€29 a month including VAT billed annually, or €348 a year).

Sources

• ICO, What is special category data?.
• Pearson Clinical UK, Telepractice and the BDI-II.
• CNIL, RGPD et professionnels de santé libéraux (in French, 1 June 2018).
• Code de déontologie des psychologues 2021, CERéDéPsy (in French, consolidated 9 September 2021), article 23.
• Data Protection Commission (Ireland), Special Category Data.