Skip to content

v0.1.2 — First Public Release

Choose a tag to compare

View all tags
@irfanalidv irfanalidv released this 03 Apr 06:10
· 5 commits to main since this release
v0.1.2
7b836a2

AgentCare v0.1.2 — First Public Release 🏥

AgentCare is a library-first Python framework for building voice AI workflows
in healthcare and care coordination. A real call comes in → data is extracted →
missing fields recovered from memory → appointment booked → confirmation email
sent → operations dashboard updated. All in one orchestrated flow.

1,000+ downloads on PyPI · 359/month · beta

✨ What's included

Voice AI Workflow Engine

  • frontdesk_booking — end-to-end call intake to confirmed appointment
  • care_navigation — route patient to right doctor/specialty from call reason
  • followup_outreach — post-visit follow-up coordination
  • Policy-based triage, conflict detection, and automation scoring

Call Intelligence

  • Transcript extraction via Mistral LLM
  • RAG fallback memory search — recovers missing patient data automatically
  • Doctor/specialty assignment from reason + policy rules
  • Idempotency tracking — no duplicate bookings from webhook retries

Appointment Connectors

  • Cal.com integration (CAL_API_KEY) for real booking
  • Mock EHR service for local dev/demos (no external deps)
  • Slot availability check before confirming

Confirmation & Notifications

  • Professional patient confirmation emails via Resend
  • Customizable sender identity

Operations Dashboard (Plotly Dash, port 8050)

  • Call Console — place outbound calls, monitor lifecycle
  • Workflow Status — health of all services live
  • Operational Case Queue — triage, conflict flags, recommended actions
  • Recent Executions — outcomes and costs
  • Appointments + Visit Details — scheduled entries with doctor and purpose
  • Auto-refreshing — no manual reload needed

Persistence

  • JSON store (local dev, zero config)
  • Postgres/Supabase (production)
  • auto mode selects the right backend based on env

Services (all FastAPI)

  • webhooks — call event ingestion (port 8030)
  • analytics — metrics API (port 8040)
  • llm_gateway — local OpenAI-compatible LLM routing (port 8010)
  • mock_ehr — local scheduling mock (port 8020)
  • dashboard — Dash UI (port 8050)

⚡ Quick Start




  

        

          
    

          
      Assets
      2
    

      
      
  
        
      Loading