Call Auto-Agent

Zero-fluff starter to:

• Answer inbound calls
• Ask 7 deterministic questions via speech
• Save a clean Markdown note to data/
• Schedule an outbound follow-up call at the captured time
• On outbound, ask for the outcome and append it to the same .md

Quickstart

1. Twilio

• Buy/verify a Voice number
• Console → Phone Numbers → your number → Voice webhook:
  • Request URL: {BASE_URL}/voice/incoming
  • HTTP: POST

2. Local run

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env  # fill values
python app.py

Tunnel (choose one):

# ngrok example
ngrok http 5000
# then set BASE_URL to the https URL ngrok prints

3. Test

• Call your Twilio number, answer the 7 prompts. Speak clearly.
• Find the Markdown in data/.
• For follow-up, say a near-future time like 2025-08-19 17:10 (24h clock).
• When the time hits, you'll get an automated call asking for the Outcome; it appends to the same .md.

Notes

• Uses Twilio speech recognition via <Gather input="speech">.
• Uses APScheduler for single-shot outbound scheduling.
• OpenAI + Notion are optional.
• For production, persist CALL_STATE in Redis/Postgres and deploy on a public server (Render/Cloud Run/EC2).

— Generated 2025-08-19T09:47:36.084875