voiceloop: the fastest voice agent loop in the browser is now open source
Everybody can now build the best voice agent into their own product. The #1 loop is open source, and the benchmark that says so is public. Demo, no keys: todoforai.github.io/voiceloop Library: github.com/todoforai/voiceloop — npm i @todoforai/voiceloop Bench: github.com/todoforai/voice-agent-bench Why we built it We wanted the fluid JARVIS feel in the browser for TODOforAI — you talk, it answers…
A new open source JavaScript library called voiceloop allows developers to easily add a voice agent to their own products. The benchmark for this library is publicly available at github.com/todoforai/voice-agent-bench. The library provides a zero-dependency setup for the full voice loop in the browser, including voice activity detection (VAD), speech-to-text (STT), large language model (LLM) processing, and text-to-speech (TTS).
Key features include real barge-in capability, self-echo filtering, and local-first implementation using WebAssembly. The library is pluggable, supporting any OpenAI-compatible LLM, four STT providers, and multiple TTS options. A JavaScript code snippet demonstrates how to integrate the library into a product, with a start() function called when the user unlocks audio.
The voiceloop library aims to provide fast, reliable voice agent functionality comparable to closed APIs, with transparent benchmark data showing it performs well in various scenarios.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
