Speech recognition has been missing half the conversation. Soniox STT v6 is the world’s first STT model that knows what your voice agent is saying. Introducing agent text input. Developers can stream
The Hook
“Voice agents break fast if they mishear a word. 'I'll' can quickly become 'Aisle'.”
Product Reveal
“With Agent Text Input, v6 understands both sides of the conversation.”
Feature Teaser
“Simply stream the LLM output in real time so the STT stays aware of what the agent is saying.”
Problem Agitation
“Done. Aisle seat it is.”
Call to Action
“Use STT v6 today.”
Synthesizable in Remotion · key components of the ai-speech-agent-showcase template.
Dashed oval trajectory with a glowing pill traversing continuously along the path
SVG path with strokeDasharray and strokeDashoffset driven by frame count, animating a moving circle along path via getPointAtLength
Floating rounded rectangle cards that pop, slide over incorrect text, and cross-dissolve
Spring-based transform scale and translateY using Remotion spring with opacity crossfade
Purpose. Demonstrate the fragility of speech recognition homophones.
Execution. Chat bubbles highlighting 'I'll' mishearing.
Purpose. Introduce bidirectional contextual awareness.
Execution. Architecture block diagram with feedback loop animation.
Purpose. Drive conversion and enterprise confidence.
Execution. Cinematic nature aerials overlaying clear brand URL.