SIPHON

👋 Hey Product Hunt! We are the creator of Siphon, and we're excited to share what we've built. 🤔 The Problem: Building AI calling agents are brutally complex. You need to: • Wrangle SIP trunks and telephony providers • Manage real-time audio streams with sub-500ms latency • Orchestrate LLM, STT, and TTS in perfect sync • Handle interruptions, state management, and call persistence Teams spend weeks just on infrastructure before writing a single line of conversation logic. ✨ The Solution: Siphon is a production-ready Python framework that abstracts all of this away. Here's a complete AI receptionist: from siphon.agent import Agent from siphon.plugins import openai, cartesia, deepgram agent = Agent( agent_name="receptionist", llm=openai.LLM(), tts=cartesia.TTS(), stt=deepgram.STT(), system_instructions="You are a helpful receptionist..." ) agent.start() That's it. 🎉 🔥 What Makes It Different: • Telephony-First: Built for real phone calls via SIP (Twilio, Telnyx, etc.) • Plug & Play: Swap providers with a config change—no vendor lock-in • Production-Ready: Auto-scales horizontally, persists transcripts/recordings/metadata • Open Source: Apache License 2.0 🎯 Perfect For: • Customer support automation • Appointment scheduling • Sales qualification • Any other use case where you need an calling agent 🙏 What We Need From You: We're launching today to get feedback from the developer community. Try it out, break it, tell us what's missing. We want to make this the default way people build telephony/calling agents. Questions? Drop them below—I'll be here all day! 👇 🔗 Links: • GitHub: https://github.com/blackdwarftec... • Docs: https://siphon.blackdwarf.in/docs • Try it: pip install siphon-ai