medtests.de · WebMCP

Evidence-based screenings, usable by AI agents

Four native browser tools let an agent select a screening, ask the exact questions, score only after explicit consent, and switch to crisis resources when necessary. German and English are supported.

Available tools

medtests_list_screenings
Find an appropriate evidence-based screening without inferring a diagnosis.
medtests_get_screening
Retrieve the exact published questions and permitted answer options.
medtests_score_screening
Score a complete answer set only after explicit consent, without storing answers.
medtests_get_crisis_resources
Stop the screening flow and return crisis contacts when urgent help may be needed.

Try it with a browser agent

Open medtests.de and use its WebMCP tools. List the available screenings in English, retrieve the GAD-7 questions, and explain the privacy and consent rules before asking any question. Do not infer a diagnosis.

The tools register automatically through document.modelContext. The same capabilities are available through the public MCP endpoint and the machine-readable API.

Start with your own AI chats

Use the guided chat-analysis prompt to let an agent review only the conversations you approve — from the current chat to selected exports from ChatGPT, Codex, Claude, Gemini, Cursor, or other tools.

Safety and privacy contract

Screening indication only, never a diagnosis.
Explicit consent is required before scoring.
Answers are neither stored nor returned.
Published wording and answer scales remain unchanged.
Crisis signals take precedence over completing a questionnaire.

Deutsch: Die Tools liefern ausschließlich Orientierung, verlangen vor der Auswertung eine ausdrückliche Einwilligung und speichern keine Antworten.