@marts
AI dev + engineer
Nothing here yet.
Nothing here yet.
1d ago · 11 min read · Your speech-to-text model has a great word error rate. Your benchmarks look solid. So why are users still complaining that the transcription "feels wrong"? Because WER doesn't measure what customers c
Join discussionMay 19 · 11 min read · Building multilingual voice agents requires coordinating four critical components—speech-to-text, language models, text-to-speech, and orchestration software—all working together within strict timing
Join discussionMay 7 · 7 min read · What "raw" means here 5-minute quickstartRaw WebSocket (this tutorial) Lines of Python~80~210 Events handled6All 13 (every server event named) Partial transcripts (transcript.user.delta)✖✔ live, overwriting line Tool calling✖✔ get_w...
Join discussionMay 7 · 12 min read · Vibe coding a voice agent used to mean wrangling three separate APIs—one for speech-to-text, one for the LLM, one for text-to-speech—and hoping your AI coding assistant could hold the glue logic together across all three. That's changed. AssemblyAI's...
Join discussionApr 29 · 13 min read · A recent survey of voice agent builders found that 82% feel confident in their technology—but 55% of end users are frustrated with interruptions, mishearing, and repetition. That gap between builder confidence and user reality gets wider in noisy env...
Join discussion