Pipecat
Framework for real-time voice agents: connects telephony, speech-to-text, an LLM and text-to-speech.
AI phone receptionist: answers calls, books appointments into the calendar, hands off to a human.
Clear pain (missed calls = lost jobs) and recurring revenue. Real work is telephony, latency tuning and call-quality testing, not the install.
BSD-2-Clause (LICENSE file read; copyright 2024-2026 Daily)
Yes under BSD-2-Clause: no restriction on building or hosting voice agents for paying clients.
Not archived; last push 2026-10-07. Maintained by Daily. It is a framework, not a finished product: every agent needs paid third-party STT, LLM and TTS services (billed by those providers) plus a transport/telephony provider. No GPU needed unless you self-host models.
Clinics, trades, salons, restaurants, any business that misses calls
AI phone receptionist: answers calls, books appointments into the calendar, hands off to a human.
Yes - Pipecat Cloud, run by Daily (usage-based, no plan names): agent-1x $0.01 per minute active / $0.0005 per minute reserved; agent-2x $0.02 / $0.0010; agent-3x $0.03 / $0.0015; Daily WebRTC voice+video $0.004 per participant minute; PSTN dial-in/out $0.018 per minute; SIP $0.003-$0.02 per minute; recording audio-only $0.005 per minute. Pipecat Enterprise (deployed in your own cloud) and enterprise support: contact sales.
None named