DocPilot aims to overcome the limitations of traditional Electronic Medical Record (EMR) systems, which are frequently cumbersome, costly, and challenging for healthcare providers to navigate. By utilizing advanced AI for transcription and data extraction, DocPilot eliminates the need for manual data entry, streamlining record management and enhancing data accuracy. This innovative approach addresses the inefficiencies of conventional EMR systems, promoting a more effective healthcare delivery system.
- 🎯 Flutter & Dart
- 🗄️ Appwrite
- 🤖 Google Gemini API (gen-ai)
- 🗣️ Deepgram’s
nova-3-medicalmodel API (leading in medical transcription, speech-to-text)
| Dashboard | Calendar | Manage Patient |
|---|---|---|
 |
 |
 |
| Patients | Prescribe | Generated Prescription |
|---|---|---|
 |
 |
 |
| Login Screen (Doctor) | Confirmation Code Screen | Login Screen (Patient) |
|---|---|---|
 |
|
 |
Check out our app's interface on Figma: Figma Design Link
A brief overview of the folder structure is defined in
Our detailed step-by-step guide will help you get up and running:
DocPilot addresses the challenges of legacy EMR systems by radically simplifying the process for medical professionals:
By capturing the entire consultation audio, DocPilot utilizes cutting-edge speech-to-text and language model APIs to reliably transcribe and extract critical medical data in real time. This removes the need for manual data entry and minimizes errors.
Instead of navigating multiple disparate systems, doctors access a single application that automatically generates structured prescriptions from their conversations. This leads to:
- 📝 Reduced Administrative Burden: No more time wasted on manual transcription and record keeping.
- ✅ Improved Accuracy: Automatic extraction and prompt correction capabilities ensure critical medical details are correctly recorded.
- ⏱️ Faster Patient Turnaround: With a digital-first approach, prescriptions can be reviewed, signed, and shared almost instantaneously.
Incorporating scheduling capabilities via a calendar module lets doctors manage consultations and follow-ups from within the same ecosystem. This end-to-end integration increases efficiency and ensures continuity of care.
With robust authentication, digital PDF signing (using individual PFX certificates), and secure storage practices, DocPilot is built to meet data security standards and instill trust among its users.
- Uses Deepgram’s medical transcription API to convert recorded consultations into text.
- Leverages Google GenAI (e.g., gemini-pro) to parse the transcription and generate a structured JSON prescription that includes symptoms, diagnosis, medications (with dosage and frequency), tests, and follow-up instructions.
- Upload: The doctor’s audio consultation is uploaded and stored securely using Appwrite’s storage.
- Processing: An Appwrite function orchestrates the transcription and AI extraction steps.
- Output: The structured output is rendered as an editable prescription on the doctor’s dashboard.
- Integration with the
syncfusion_flutter_calendarenables real-time scheduling, displays upcoming appointments on the doctor’s dashboard, and allows patients to book consultations directly.
- Designed in Flutter with a clean, responsive UI.
- Uses state management (BLoC) and dependency injection (GetIt) under a Clean Architecture paradigm to ensure maintainability and scalability.
- The app generates PDF prescriptions using
syncfusion_flutter_pdf. - Implements a digital signing mechanism using individual PFX certificates generated per doctor, ensuring the prescription integrity and authenticity, backed by biometric verification on the client side.
The doctor initiates the process by recording and uploading the consultation audio, which is stored on Appwrite and generates a publicly accessible URL.
An Appwrite function calls Deepgram’s nova-3-medical model to convert audio into text, ensuring high accuracy in medical terminology.
The transcription is then sent to Google GenAI with a prompt that instructs the model to extract relevant medical data and output a structured JSON.
The app parses the JSON, allowing the doctor to review the prescription. Once approved, the prescription is converted into a PDF and digitally signed with the doctor’s PFX certificate.
Patients and doctors manage and schedule appointments seamlessly through an integrated calendar view.