September 7, 2026 · release

@voicethere/agent

VoiceThere can now identify the language the caller is speaking and deliver a user_language event to your agent. Use the ISO 639-1 code to switch prompts, pick a TTS voice, or keep session state in the caller’s language.

Detection runs on-cluster with Sherpa Whisper tiny. The first confident code is emitted once; it is re-sent only if the language changes later in the session. Language events are not stored as conversation-history turns.

Try it

Docs: Spoken language identification
Dashboard: VoiceThere dashboard → your project → Voice (STT / TTS)
Agent: handle onUserLanguage in @voicethere/agent

import { defineAgent, agentLog } from "@voicethere/agent";

export default defineAgent({
  onUserLanguage({ sessionId, language }) {
    agentLog("info", `detected ${language} for ${sessionId}`);
  },
});

Spoken language identification for voice agents

Try it

Similar posts