Spaces:
Running
on
Zero
Running
on
Zero
Y Phung Nguyen
commited on
Commit
·
3b38a6c
1
Parent(s):
ef322a1
Upd README
Browse files
README.md
CHANGED
|
@@ -53,6 +53,12 @@ tags:
|
|
| 53 |
- Responses translated back to original language
|
| 54 |
- Powered by Gemini MCP for translation
|
| 55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
### 🎤 **Voice Features**
|
| 57 |
- **Speech-to-Text**: Voice input transcription using Gemini MCP
|
| 58 |
- **Inline Mic Experience**: Built-in microphone widget with live recording timer that drops transcripts straight into the chat box
|
|
|
|
| 53 |
- Responses translated back to original language
|
| 54 |
- Powered by Gemini MCP for translation
|
| 55 |
|
| 56 |
+
### 🧾 **Clinical Intake Q&A Breakdown**
|
| 57 |
+
- Gemini intake triage checks whether the user’s concern needs additional questioning (up to 5 follow-ups) and keeps per-session state.
|
| 58 |
+
- Intake agent conducts focused Q&A, then auto-summarizes the transcript into patient profile, refined problem statement, and key findings with actionable handoff notes.
|
| 59 |
+
- Pipeline injects both the structured insights and raw transcript back into the supervisor so downstream planning, RAG, and search stay grounded in what the patient actually said.
|
| 60 |
+
- Users can toggle the intake flow directly in the UI; disabling clears any pending follow-up state.
|
| 61 |
+
|
| 62 |
### 🎤 **Voice Features**
|
| 63 |
- **Speech-to-Text**: Voice input transcription using Gemini MCP
|
| 64 |
- **Inline Mic Experience**: Built-in microphone widget with live recording timer that drops transcripts straight into the chat box
|