Phlox is an open-source patient management system integrating AI-powered medical transcription, clinical note generation, and an AI chatbot interface. It's designed to run locally, utilizing Ollama and Whisper for inference and transcription.
- 🔒 100% Local & Private: Runs entirely on your machine with no third-party services - all data stays local, using only free, open-source tools.
- 🎤 AI Medical Transcription & Summarization: Convert audio to structured clinical notes using customizable templates.
- 📝 Flexible Template System: Structure clinical notes to your preferences, with versioning and automated template generation from example notes.
- ✅ Task Manager: Parse clinical plans into actionable task lists with AI-generated summaries.
- ✉️ Correspondence Generation: One-click generation of patient letters based on clinical notes.
- 🤖 AI-chat/RAG: Chat with an LLM about cases, backed by a local document knowledge base (ChromaDB).
- 🧠 Clinical Reasoning: AI-assisted differential diagnosis and investigation planning
- 📰 Dashboard with RSS Reader: Stay updated with LLM-summarized articles from medical RSS feeds.
- 🆓 Free & Libre Software: Completely free (as in freedom) - study, modify, and share as you wish.
- Frontend: React/Chakra UI
- Backend: FastAPI
- Database: SQLite
- LLM Backend: Ollama
- Transcription: Whisper
- RAG: ChromaDB
- Prerequisites: Podman/Docker, Ollama, Whisper endpoint.
- Hardware Requirements: For reasonable performance, a GPU (CUDA, ROCm) or Apple M-Series chip is strongly recommended. Without these, especially with larger models, the system will run extremely slowly.
- Clone:
git clone https://github.com/bloodworks-io/phlox.git && cd phlox
- Build:
docker build -t phlox:latest .
- Environment: Create
.env
inphlox/
(see example in documentation). - Run:
docker-compose up
(Production) ordocker-compose -f docker-compose.dev.yml up
(Development). - Access: http://localhost:5000
For detailed setup, feature explanations, and important warnings, please see the Documentation.
Here's what's coming next for Phlox:
- Use structured JSON outputs for managing LLM responses
- Add support for OpenAI-compatible endpoints
- Create Electron app packaging for desktop use
- Develop Progressive Web App (PWA) version
- Introduce advanced template version control
- Meeting and multi-disciplinary meeting scribing
Phlox is an experimental project intended for educational and personal use. It is not a certified medical device and should NOT be used for clinical decision-making without thorough validation, regulatory approvals, and under the direct supervision of qualified medical professionals.
Key limitations:
- Experimental Code: The codebase is a work in progress and may contain bugs and inconsistencies.
- AI Hallucinations: LLM outputs, especially from smaller models, can be unreliable, inaccurate, and may present plausible but incorrect information. Always verify AI-generated content against trusted sources and use your professional clinical judgment.
- No User Authentication: Naively exposing this application to the open internet is highly discouraged. Phlox has no user access controls and, for now, next to no input sanitisation.
- Not HIPAA/GDPR Compliant: Phlox, in the form provided in this repo, lacks the necessary security and compliance measures for handling protected health information in regulated environments.
Use at your own risk and only for non-clinical, educational purposes unless you have implemented robust security measures and undertaken thorough validation.