All answers about Psynex
at a glance
All answers about Psynex: AI documentation, dictation software, context chat, and more.
14 days free • No credit card • GDPR compliant
Conventional dictation software like Dragon or Windows speech recognition only transcribes spoken text. Psynex goes further: It automatically creates structured therapy documentation from the transcript, recognizes ICD-10 diagnoses, tracks symptoms across sessions. You don't dictate – you document automatically.
Yes. Our AI-based speech recognition has been trained with various speech patterns and reliably understands regional accents. Recognition accuracy is over 95%, even at faster speaking speeds.
Yes, Psynex was designed exactly for this. You start the recording at the beginning of the session, conduct your conversation as usual, and Psynex automatically creates the transcription afterwards. The microphone records discreetly – the therapeutic relationship remains undisturbed.
Audio files are used exclusively for transcription and are automatically deleted immediately afterwards. No audio recordings are stored. Only the text transcript and the documents created from it remain in your account.
No. Psynex works with any microphone – including the built-in laptop or tablet microphone. For optimal results in larger rooms, we recommend an external microphone, but it's not a requirement.
Our speech recognition is optimized for psychotherapeutic and psychiatric vocabulary. ICD-10 codes, therapy methods (CBT, DBT, EMDR, etc.), medication names, and clinical terms are reliably recognized.
Yes, you must inform your patients about the recording and processing and obtain their consent. Psynex provides you with a template for patient information that you can use directly.
The AI analyses serve as support and guidance – not as a replacement for your clinical judgment. Accuracy varies depending on conversation content and quality. The key point: Every assessment is backed by direct quotes and timestamps, so you can verify everything. We recommend always critically reviewing the AI suggestions.
No, never. Your patient data is used exclusively for the analyses you request. No training of AI models takes place with your data. This is contractually established in our Data Processing Agreement and in the BAA with OpenAI.
The AI analyzes the transcript for linguistic indicators of suicidality and categorizes them on a 5-level scale (0-4). Each assessment is backed by the exact quote and timestamp. Important: The screening is a tool for attention guidance, not a diagnostic instrument. The clinical judgment always lies with you.
Psynex specializes in F-diagnoses (Mental and Behavioral Disorders) of the ICD-10. The AI recognizes indicators of diagnoses and suggests primary, secondary, and differential diagnoses – always with source references from the conversation. The final diagnosis naturally remains your medical/therapeutic decision.
Yes. You can create your own document templates and instruct the AI to document according to your specifications. This way, Psynex adapts to your individual documentation style and the requirements of your practice.
In the context chat, you can start a chat session and link it to a patient and selected sessions. The AI then has access to this session data and can answer your questions contextually – e.g., about symptom development over time, therapy progress, or preparation for supervision presentations.
No, absolutely not. Psynex is a documentation and analysis tool that supports you – not a replacement for your expertise. All AI analyses are suggestions that must be reviewed by you. You make the therapeutic and diagnostic decisions. We show the notice at every point: 'AI can make mistakes. Please verify important information.'
The AI analyzes the spoken content for indicators of the five main symptom areas (anxiety, depression, sleep problems, lack of energy, tension) and estimates the intensity on a 0-10 scale. Each rating is linked to the relevant conversation passages and timestamps so you can understand the assessment.
The symptom ratings are based on analyzing linguistic indicators in the conversation. They are meant as guidance, not as clinically validated measurement instruments. The strength lies in comparability over time: When you use the same AI methodology, you can recognize relative changes well. Each rating is backed by quotes, so you can understand the assessment and correct it if needed.
Psynex specializes in the F-diagnoses (Mental and Behavioral Disorders) of ICD-10, including: Affective disorders (F30-F39), Anxiety disorders (F40-F41), Stress disorders (F43), Somatoform disorders (F45), Personality disorders (F60-F69), and others. The AI recognizes indicators and suggests diagnoses – the final diagnosis remains your medical/therapeutic decision.
The screening is an attention tool, not a replacement for your clinical judgment. It analyzes linguistic cues and alerts you to potential signs that you then evaluate yourself. It can miss or misinterpret indicators. Use it as a supplementary information source, not as the sole basis for clinical decisions. In cases of acute suicidality, your established clinical protocols always apply.
'High' means that several clear indicators were found in the conversation. 'Medium' shows some hints, but less clear. 'Check' for differential diagnoses means that the AI sees alternative explanations for the symptoms that you should consider. These assessments are based on the quantity and clarity of linguistic indicators – not on clinical validation.
The AI analyzes the entire transcript and identifies recurring themes, emotionally charged passages, and explicitly named problem areas. It extracts up to 5 main topics per session and links each topic to the relevant conversation passages. This gives you a quick overview and allows you to dive deeper if needed.
No. Your patient data and analyses are used exclusively for the evaluations you request. No training of AI models takes place with your data. This is contractually established in our Data Processing Agreement and in the BAA with OpenAI.
The five symptom categories (anxiety, depression, sleep problems, lack of energy, tension) are currently fixed, as they cover the most common areas in psychotherapy. We are working on adding individual categories in the future. For specific analyses, you can use the context chat and ask targeted questions.
Psynex saves the analysis results of each session and can aggregate them over time. You see trends, changes, and can recognize patterns. The progress display shows you at a glance how individual symptoms have developed – helpful for therapy planning, supervision, and reports.
You can link any number of sessions to a chat – from a single one to all sessions of a patient. The more context the AI has, the better it can recognize progressions and patterns. With many sessions, we recommend selecting the relevant ones specifically for more precise answers.
Yes. When you link sessions to the chat, the AI has access to the complete transcripts and analyses of those sessions. This allows it to provide concrete quotes and refer to specific conversation moments.
Yes, chat histories are saved in your account so you can access them later. You can delete any chat at any time. All data is encrypted and not used for AI training.
Absolutely! The Context Chat is excellent for preparing supervisions. Ask e.g. 'Summarize the therapy progress so far for supervision' or 'What were the key developments of the last 5 sessions?' – and you'll get a structured overview.
The AI bases its answers on the actual session content. It quotes directly from transcripts and analyses. Nevertheless: AI can make mistakes or misinterpret connections. Use the answers as guidance and verify important information.
Yes. You can also ask general questions, e.g. about therapeutic techniques, ICD-10 criteria, or documentation requirements. The chat also works without a linked patient – then the AI simply doesn't have access to your specific session data.
In the Premium plan, there is no limit on chat messages. In the Basic plan, the number of monthly messages is limited. Details can be found on our pricing page.
Yes. You can export chat histories as text, e.g. to include a summary in your documentation or use for supervision materials.
Psynex's OCR technology uses modern AI models that reliably recognize even difficult handwriting. Recognition accuracy is over 90% for most handwriting. With very illegible writing, individual words may be incorrectly recognized – that's why there's always the option for manual correction.
You can upload photos in all common formats: JPEG, PNG, HEIC (iPhone). PDFs with scanned documents are also supported. For best results, we recommend well-lit photos with sufficient contrast.
Yes. You can upload multiple images and save them as a related document. The OCR processes each page individually and combines the recognized text.
You have the choice: You can save the original image together with the recognized text or keep only the text. The original can be helpful for checking when things are unclear.
Yes, even better! Printed text is recognized with even higher accuracy than handwriting. So you can also digitize medical letters, reports, or other printed documents.
Yes. After text recognition, you can fully edit, correct, and supplement the recognized text. Only then is the document saved and optionally assigned to a patient.
Digitized notes are stored as documents in your Psynex account. You can assign them to a patient or file them as general notes. All data is encrypted and stored on European servers.
Yes. Once a note is digitized and saved, it can be used by other Psynex features – e.g., in context chat or for reports. This way, older handwritten records are also integrated into your digital workflow.
Still have questions?
Our support team is happy to help. Contact us or try Psynex free for 14 days.