MongoDB - Build AI That Scales
Mini toolsInk 2
By Cartesia
Ink 2 is Cartesia’s transcription model for production voice-agent systems. It is built for real-world streaming speech-to-text in environments with telephony audio, background noise, varied accents, structured data, and conversational pauses. Ink 2 natively handles phone numbers, dates, emails, currencies, UUIDs, and other structured speech, and it includes model-level turn detection with turn.start, turn.end, and turn.eager_end signals. Cartesia describes it as using semantic endpointing, meaning it detects turn completion by meaning rather than only silence, reducing premature interruptions in live agent conversations.
New
Multimodal
Gen 3
Overview
Ink 2 is Cartesia’s speech-to-text model for voice agents, optimized for streaming transcription, semantic endpointing, noisy audio, and low-latency turn detection.
About Cartesia
Industry:
Artificial Intelligence
Company Size:
91
Location:
Daly City, California, US
Website:
cartesia.ai
Tools using Ink 2
No tools found for this model yet.
Last updated: June 16, 2026
