Tag
A Kody package that enables any Kody-connected AI agent to make outbound phone calls via Bland.ai, handling the call lifecycle and transcript reading.
OpenAI has launched GPT-Live-1, a full-duplex voice model for API that enables natural, bidirectional voice conversations for developers, reducing latency and improving turn-taking in voice agents.
The author shares a postmortem on building a production phone-based AI voice agent, revealing that most engineering time was consumed by telephony infrastructure, turn detection, observability, and failure handling rather than core LLM behavior. They suggest using managed platforms like Vapi, Retell, or Dasha from the start to focus engineering effort on business logic.
A technical breakdown of an enterprise Voice AI architecture that cuts telephony costs by 40-60% via wholesale carriers and achieves sub-500ms latency using Deepgram, Claude/GPT-4o-mini, and ElevenLabs/Cartesia, orchestrated through n8n and Supabase.
An analysis of 10,000 voice AI calls reveals key failure points: STT error rates, first-8-second chaos, interruption handling, extended silence, tool call latency, LLM failures, and broken escalation — offering practical insights for agent builders.
The author built and open-sourced a telephony layer for AI agents, enabling them to make calls and send texts via a REST API and MCP server, designed to work with frameworks like OpenClaw, Hermes, LangChain, and Claude Code.
Open-sourced a full-stack production starter for voice agents using LiveKit, FastAPI, and React, handling both web and telephony with a single code path, deployable via Docker Compose.
A detailed explainer on the five-layer architecture of AI voice agents, including speech-to-text, LLM, text-to-speech, orchestrator, and telephony, all operating under a 500ms latency constraint to maintain natural conversation flow.