{
  "id": 13029390,
  "title": "Shaving Latency Off Real-Time Speech Translation: What Actually Worked in Our Flutter App",
  "url": "https://urgent.news/2026/10/09/shaving-latency-off-real-time-speech-translation-what-actually-worked",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-10-09T04:39:30.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/mrzhangguoguo/shaving-latency-off-real-time-speech-translation-what-actually-worked-in-our-flutter-app-2kff"
  },
  "original_language": "en",
  "account": null,
  "summary": "The article discusses the challenges and solutions in reducing latency in a real-time speech translation app called Owll Translator. The app uses earbuds to translate speech in real-time, with the option to speak the translation in the user's own voice. The author explains that most of the latency work focused on hiding unavoidable waits, cutting down avoidable delays, and getting sound into the listener's ear as soon as possible. The article details the two main backend paths used in the app - client-orchestrated and server-orchestrated - and outlines various techniques employed to minimize latency, such as starting the translation process before the user initiates it, prefetching authentication tokens and creating sessions in advance, and optimizing the audio pipeline. The author also acknowledges the lack of reliable end-to-end latency measurements and explains the steps being taken to address this issue.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}