Workflow, use cases, best practices, and frequently asked questions — everything you need to mentalhealthGPT the most out of mentalhealthGPT .
From the first case file to AI-assisted work in a real-world clinical setting.
Erstellen Sie für jeden Klienten ein eigenes Dossier. Das Dossier ist der zentrale Arbeitsbereich — hier werden alle Sitzungen, Dokumente und Notizen gebündelt. Name und Inhalt werden verschlüsselt gespeichert und sind nur in Ihrem Konto sichtbar.
Detailed instructions →Nehmen Sie die Sitzung direkt im Browser auf oder laden Sie eine bestehende Audiodatei hoch. mentalhealthGPT transkribiert die Aufnahme und erstellt automatisch strukturierte Dokumentation. Die Spracherkennung erfolgt lokal im Browser; Sitzungsaufnahmen werden nicht auf unseren Servern gespeichert.
Detailed instructions →Wireless wearable microphones are highly recommended — they are the only way to ensure accurate transcription and reliable speaker separation. Anything the microphone fails to capture clearly enough cannot be reconstructed by AI later on.
Detailed instructions →Create reports directly from the file. Upload existing documents and add professional notes. The more complete the file, the more accurate the AI-powered assistance.
Detailed instructions →Ask the AI specific questions — based on the context of this dossier, your context library, and the model’s extensive knowledge. No generic answers: every response is based on your actual session data.
Detailed instructions →mentalhealthGPT you throughout the various stages of your clinical work.
Review transcripts, summarize key points, update case documentation — all in a fraction of the time, without losing any information.
Select relevant excerpts from sessions, organize case materials, and formulate questions — directly from the case file, pseudonymized and focused.
Over the course of weeks and months , a clear picture of the course of therapy emerges — themes, patterns, and changes become apparent.
Upload existing reports, assessments, and preliminary documentation and use them as context for the AI — no more manual summarization.
These six habits make the biggest difference in daily use.
Wireless wearable microphones are highly recommended — they are the only way to ensure accurate transcription and reliable speaker separation. AI cannot reconstruct audio that the microphone does not capture clearly enough. Recommended devices and setup →
Even a 20-minute recording provides valuable data. The AI needs cumulative session data to recognize patterns — individual sessions are effective, but the full value is realized over time.
Quickly skim through the minutes of the last meeting, refresh your memory on pending issues — two minutes of preparation with the dossier instead of twenty minutes poring over files.
Upload your therapeutic frameworks, protocols, and guidelines. The AI will then respond using your language and methodology — not in a generic way.
Anything the transcript doesn't capture — body language, mood, clinical impression — should be included in the file as a brief note. The text and audio complement each other.
Instead of asking, “What could I do?” try asking, “What interventions fit the patterns described in the last three sessions?” — the more specific the question, the more precise the answer.
Before incorporating AI-generated responses into reports or supervision sessions, make sure they do not contain any real client data. The system automatically pseudonymizes the data — a quick glance is all it takes.
Practical answers for everyday life — not a marketing FAQ.
No, but it’s recommended. The more session audio included in the file, the more accurately the AI can reflect the actual course of events. Individual sessions work — but the full value is realized across multiple sessions.
MP3, M4A, WAV, WebM, OGG, FLAC und AAC. Höchstens 300 MB und 2 Stunden pro Datei. Die direkte Aufnahme im Browser erzeugt automatisch das richtige Format. Für Zoom-Aufnahmen steht eine dedizierte Integration zur Verfügung.
Sitzungsaufnahmen werden nicht auf unseren Servern gespeichert. Nur das verschlüsselte Texttranskript verbleibt im Dossier. Bei direkter Aufnahme und Datei-Upload erfolgt die Spracherkennung lokal im Browser; bei Zoom-Sitzungen wird der Audiostream über unsere Infrastruktur an Ihren Browser weitergeleitet und dort transkribiert.
The transcription itself is quick: typically 1–3 minutes for a 50-minute session. The subsequent speaker recognition runs as a separate step and can take an additional 10–15 minutes — the model analyzes the entire recording, and the speed depends on the device. Both steps run in the background, and the app remains fully usable during this time.
Yes. In the Documents section, you can upload PDFs, Word files, and text documents. These are indexed and made available to the AI as context — ideal for preliminary documentation, reports, and assessments.
Ja. Dossierinhalte werden verschlüsselt in der Schweiz gespeichert; für KI-Anfragen werden pseudonymisierte Inhalte an Modelle in der EU beziehungsweise der Schweiz übermittelt. Verschlüsselung, Pseudonymisierung und eine kontrollierte, protokollierte Verarbeitung schützen Ihre Inhalte. Weitere Informationen: Datenschutz & Sicherheit.
mentalhealthGPT direct recording of Zoom sessions via RTMS integration.