RagLeapShort answer: RagLeap Core has a built-in healthcare_intake role, and the code treats it as a sensitive domain. You can run it yourself, upload your clinic information, and let it answer routine questions and collect intake details, but it is not a medical device, it gives no medical advice, and it has no compliance certification. This page describes what the code does and what you must check yourself. It is not medical, legal or compliance advice.
Which role applies, and how is it treated?
The role is healthcare_intake. In the code (sensitivity.py at v0.11.0) it is one of the built-in sensitive-domain roles, along with roles such as legal intake, insurance, compliance, veterinary intake and tax preparation intake. Sensitive roles are forced to semi or off autonomy, and get a reasoning mode and a grounding check. The README says sensitive roles always need approval. A custom role you create is also treated as sensitive when words such as medical, health, patient, clinical or pharmacy appear in its name, display name or skill tags.
What would you upload?
Information your reception answers from today: opening hours, services, location and parking, how to book, what to bring, fees you publish, and your own policies. Use the Documents tab (PDF, Word and other supported files, or a public web page by address). The assistant answers from your documents and cites them. Do not upload patient records unless you have decided that your setup, your AI provider and your legal basis allow it.
How do patients reach it?
Through WhatsApp, Telegram, Discord or a voice line (Twilio). A website chat widget is planned, not built.
What should a clinic check first?
- Where the data goes. Your documents and each conversation go to the AI provider you configure. A local Ollama model keeps these steps on your own machine, but local models on a CPU are slower and less accurate than Gemini. Check your provider's terms for health data.
- Approval. Keep autonomy at
offorsemi. Read real conversations before going live. - No medical advice. No clinical accuracy is claimed. Do not use it to triage, diagnose or advise.
- Security and access. No named users or roles yet, no independent security review yet, and the compose file ships a public default Neo4j password on localhost only (a random one per install is planned).
- Compliance. RagLeap makes no HIPAA, GDPR or other compliance claim. Whether it fits your obligations is for you and your advisers.
- Single-tenant. One deployment, one clinic, one document set.
Sources
The v0.11.0 README, the v0.11.0 release notes, sensitivity.py and the general setup guide.