Sparrow-2
A turn-taking model that knows when a voice agent should speak
Emerged finds new products worth a look, daily — what’s new today
Voice agent teams can run regression suites against simulated callers with their own model keys, instead of paying a markup on every test call.
Egma lets you build regression suites of simulated voice and text conversations for your voice agent, then review real production calls, tool calls and metrics. Bring your own STT, LLM and TTS keys, or use inference inside the platform at no markup. Use the hosted cloud or self-host with the egma CLI. The core is MIT-licensed, with an enterprise folder under separate terms. It is built by two brothers who have shipped voice agents since 2024.
More like this
Nico
Hold a key, say what you need, and a sandboxed worker on your Mac does it
VoiceMem
Open streaming memory for real-time voice agents, split into facts and feelings
Jinfer
An AI inference engine that runs inside the JVM, with no Python sidecar
JARVIS
A voice assistant that brainstorms with you, then runs and watches Claude Code
botato
Sends the AI agent you already use to sit in on your video meetings