Tendem
Mini toolsSparrow 1
By Tavus
Sparrow-1 is a specialized conversational-flow model used in Tavus’s Conversational Video Interface. Operating directly on streaming audio, it models conversational floor ownership frame by frame, handling overlaps, interruptions, hesitations and affective pauses. This lets downstream ASR, LLM and TTS components start or withhold responses at the right moment, achieving human-like turn-taking with sub-100 ms median latency across languages, and prioritizing natural dialogue rhythm over naive “respond as fast as possible” behavior.
New
Audio
Gen 4
Overview
Sparrow-1 is Tavus’s multilingual audio-native model for conversational timing, predicting who holds the floor and when to listen, wait or speak so real-time voice agents match human turn-taking instead of relying on silence thresholds.
About Tavus
Tools using Sparrow 1
-
Patrice Blanchard🙏 11 karmaSep 19, 2024@TavusThe demo of this AI zagent is pretty stunning. The agent understand what I am saying zand is replying on the spot. Very interesting. The sky seems to be the limit. But how can a small company use this kind of tool in practice.
Last updated: January 21, 2026
