Build an AI agent that makes real phone calls in under 30 minutes.
| File | Description |
|---|---|
01_quick_call.py |
Make a single AI call and get the transcript |
02_batch_caller.py |
Call a list of contacts, collect structured results |
03_inbound_handler.py |
FastAPI webhook server to handle inbound calls |
04_langchain_agent.py |
LangChain agent that decides when to call |
05_crewai_outreach.py |
CrewAI multi-agent lead qualification crew |
.env.example |
All environment variables you need |
# 1. Install Spix CLI
curl -fsSL https://spix.sh/install | sh
# 2. Authenticate (creates a free account)
spix auth login
# 3. Clone this kit
git clone https://github.com/Spix-HQ/hackathon-starter
cd hackathon-starter
# 4. Install Python deps
pip install -r requirements.txt
# 5. Copy and fill in your env vars
cp .env.example .env
# Edit .env with your keys, then:
source .envTo make calls, you need:
- An Agent plan ($20/mo) or Operator plan ($99/mo) — upgrade at app.spix.sh
- A phone number:
spix phone rent --area-code 415 - A playbook: see
01_quick_call.pyfor how to create one
python 01_quick_call.py --to +19175550123python 02_batch_caller.py --csv leads.csv --output results.jsonpython 03_inbound_handler.py
# Then expose with ngrok and create a webhook:
ngrok http 8000
spix webhook create --url https://YOUR_NGROK_URL/inbound --events call.started,call.completed,call.failed- Sales qualifier: Call a list of leads, extract "interested: true/false" and "demo_slot"
- Appointment reminder: Auto-call patients/clients to confirm appointments
- Survey agent: Call participants, ask questions, aggregate results
- Support callback: Trigger a call when a support ticket is opened
- Real estate agent: Call inquiries, qualify buyers, schedule showings
- Recruiting screener: First-pass phone screen before human interview
Questions? spix.sh/discord | docs.spix.sh