Matching: counselors from our network, suggested from your answers
For your psychological counseling, our matching suggests up to three counselors who fit your concern when you book. You answer two short questions and see the reasons for each suggestion. We developed the matching ourselves. An AI only reads along if you describe your concern in your own words and explicitly agree. Whether you use the matching is up to you. Without matching you book as before: you pick a free appointment, and your counselor follows from it.
Developed in-house
The topic catalog, the rules and the selection are our own. We did not buy a ready-made matching. If you write no text, the suggestions come from your answers and our fixed rules alone, with no AI at all.
- Only if you want to use AI
Tested language models
Only if you describe your concern yourself and agree, a language model sorts the text into our topic tree, names keywords for topics our catalog does not know yet and takes over explicit wishes about the language or gender of your counselor. One of these models is used: Claude Opus 5.5 by Anthropic or Mistral Large 4 by Mistral AI. Which one, we decide by quality, data protection and price. A new model is listed here before it reads any text.
- Only if you want to use AI
Processed in Europe
If you agree, your text goes without your name and without contact details to the provider of the model in use and is processed in data centers in Europe, in the EU or in an EFTA country. It is not stored there and not used for training. Who runs which model is shown in the table below.
The only basis is your explicit consent under the General Data Protection Regulation. Wherever AI is involved, we say so. More on the EU AI Act. Evidence from the recipients: Amazon Web Services: processing within the EU, data protection at Amazon Bedrock; Mistral AI: processing in the EU and EFTA, data processing agreement, sub-processors.
The language models in detail
Who receives your text depends on the model. For Claude Opus 5.5 by Anthropic the recipient is Amazon Web Services EMEA SARL (Luxembourg) with the service Amazon Bedrock. Processing takes place in data centers in the EU, the service does not store the text and does not use it for training. The makers of the models do not receive it. For Mistral Large 4 by Mistral AI the recipient is Mistral AI (Paris) through its EU endpoint. Mistral AI is also the maker of the model. Processing takes place in data centers in the EU and in EFTA countries (Iceland, Liechtenstein, Norway, Switzerland), where the GDPR or an adequacy decision of the EU Commission applies. Zero data retention is active for our account (confirmed on 6 October 2026): Mistral AI does not store the text, not even for abuse monitoring. Use for training is switched off.
| Model | Released | Maker | Maker based in | Recipient | Processing |
|---|---|---|---|---|---|
| Claude Opus 5.5 | 22 September 2026 | Anthropic | USA | Amazon Web Services EMEA SARL (Luxembourg) | Amazon Bedrock, several regions in the EU |
| Mistral Large 4 | 6 October 2026 | Mistral AI | France | Mistral AI (Paris) | Mistral AI directly, EU endpoint (data centers in the EU and EFTA) |
Details checked on 6 October 2026.
Two ways to your appointment
| Feature | Way 1Matching | Way 2Next free appointment |
|---|---|---|
| What you do | MatchingAnswer two questions, describe your concern if you wish | Next free appointmentPick a free appointment |
| What you see | MatchingOne to three counselors with reasons and their next appointment | Next free appointmentAll free appointments |
| Who chooses | MatchingYou choose the counselor and the appointment | Next free appointmentYou choose the time, the counselor follows from the appointment |
| AI involved | MatchingOnly if you write a text and agree | Next free appointmentNo |
What goes into the suggestions
From you: your main concern from twelve topic areas and the conversation style you would like. Optional: a more precise sub-topic, up to two further topics, your age, your wishes regarding the counselor's language and gender and a few sentences about your concern. From our counselors: their topics with particular areas of focus, their conversation style, their languages, the age groups they work with and their free appointments. They maintain this information themselves.
If you start with the info call
In the free info call you can tell us what matters to you: topic, conversation style, language and, if you wish, the gender of your counselor. Our team uses these wishes and gets a suggestion from them by the same fixed rules. Whether an AI service reads your written concern is your decision, with a separate checkbox when you book the info call. Without that checkbox the suggestion comes only from what you tell us in the call, with no AI at all. When our team assigns you, a person always decides. If the team leaves the choice open, your personal booking link shows the same choice as the direct booking: next free appointment or matching. The same applies if you later change to another counselor and the choice is open.
Cost
The matching is included in the price of every session. We did not raise prices when we introduced it. We bear the cost of the AI service ourselves. We invest in it because we are convinced that we offer the best value for money in psychological online counseling in Germany.
What the AI does and does not do
The AI only reads the text you write yourself. It sorts it into our topic tree and names keywords for what our catalog does not know yet. You see the recognized topics with the suggestions and can remove each of them. It takes over wishes about the language or the gender of your counselor that you state in it explicitly. This happens only with your explicit consent. It makes no diagnosis, does not assess how burdened you are and does not decide about you. The order of the suggestions is calculated by fixed rules that we developed ourselves. The technical term is a knowledge-based recommender: written-down expertise instead of a model that learns from the behavior of many people. If you write no text, you get suggestions without any AI, from your answers alone.
No ranking by performance
We do not rank counselors by ratings or revenue. What counts is whether their focus and way of working fit your concern and whether an appointment is free soon. The less your answers tell people apart, the more the next free appointment decides, just as on the way without suggestions.
Open criteria, internal weighting
This page lists which information goes into the suggestions. How much each criterion counts, we do not publish. We review the weighting regularly against the feedback after sessions.
How the suggestions get better
Our goal is to find the counselor who fits each concern best. That is why we evaluate how well the suggestions fitted. We only adjust the rules once there is enough feedback. People decide this, not the AI. After the first session we briefly ask you and your counselor how well it fitted. We evaluate this feedback as totals to compare the booking ways. Only with your consent do we link it with your answers from the matching, without your text and without names. This is how we learn which criteria lead to a good fit. With the same consent we count the keywords from your text, without any link to you, so that our topic catalog can grow. We read free texts in the feedback form for this only with a separate tick. We do not evaluate who counsels better or worse. Right at the suggestions you can also tell us with one click whether they were helpful. This feedback is anonymous and is about the suggestions as a whole, never about an individual counselor.
Your data
The language model receives your text without name and without contact details. The text is not used for training there. We do not store it in our record of the suggestions, only your answers, the topics found and keywords. We delete all of it 30 days after the first session. If you do not book, we delete it 30 days after the request. If you agreed to the evaluation for improvement, we delete it after eight weeks. You can withdraw your consent at any time, in the client portal or by email. The counselor you book with sees, to prepare, the conversation style you would like (only if it differs from their profile) and your preferred language (if you chose more than German alone), never your topics.
Processing takes place in Europe, in the EU or in an EFTA country. Details in the privacy policy
If it does not fit
You can change your counselor at any time. If it does not fit after a paid session, our money-back guarantee applies: 7 days, as long as you have not booked another session.
What research says about fit
For psychotherapy it is established that taking wishes into account lowers dropout. This is shown by a meta-analysis in the Journal of Clinical Psychology (Swift and colleagues, 2018) and a review in JAMA Psychiatry (Windle and colleagues, 2020). This evidence does not exist for psychological counseling in the same way. That is why we promise no effect but measure: after every first session we ask how well it fitted.
Background on the blog: how we use AI as a tool and what you gain from it
Matching at Psychofit since 6 October 2026. As of October 2026
Frequently asked questions
How do I find a psychological counselor online who fits me?
Choose "Matching" when you book. You choose your main concern from twelve topic areas and the conversation style you would like. If you wish, add a more precise sub-topic, up to two further topics, language, your age and whether you would rather talk to a woman or a man. You see up to three counselors with reasons and the next free appointment and choose yourself. If it does not fit after the first session, you can switch at any time.
Does the matching cost anything?
No. The matching is part of what we offer and is included in the price of every session. We did not raise prices when we introduced it. We bear the cost ourselves.
Do I have to use the matching?
No, it is your choice. The way via the next free appointment remains equivalent. Even after the suggestions you can view all appointments.
Does my text go to an AI?
Only if you write a text and explicitly agree to it being sent. Without that tick the suggestions come from your answers alone.
Does the AI decide who counsels me?
No. AI is only used if you describe your concern in your own words and tick the box for it. It then sorts your text into our topics so that the suggestions become more precise. The order comes from fixed rules. You make the choice.
Is my text used to train an AI?
No. Your text is not used for training. We do not store it in our record of the suggestions. The language model only receives it with your explicit consent and without your name or contact details.
Use the matching when you book
In the booking form, choose "Matching: find fitting counselors (AI only if you wish)". Without matching you book as before via the next free appointment.