Krisp for Developers
Product · API / developer platform · Sales by Krisp
Krisp for Developers provides a suite of SDKs and APIs for integrating real-time Voice AI into applications and AI agents. The VIVA SDK is specifically designed for Voice AI agents, offering models for voice isolation, turn prediction, and interruption detection to improve STT accuracy and natural conversation flow. The RTC SDK focuses on human-to-human communication, providing noise cancellation and accent conversion for contact centers and communication platforms. Additionally, the Voice Translation API offers self-serve, real-time speech-to-speech translation across 61 languages. These mode
- Price
- From 400 USD month
- Languages
- 61 languages for translation
- Latency
- 15 ms for Voice Isolation
- Audio Format
- PCM S16LE, 16 kHz mono
- Supported Platforms
- Windows, Mac, Linux, Web (JS/WASM), iOS, Android
- CPU Optimization
- AWS Graviton3 (Arm NEON), Intel Sapphire Rapids (AVX-512)
Overview
Krisp for Developers provides a suite of SDKs and APIs for integrating real-time Voice AI into applications and AI agents. The VIVA SDK is specifically designed for Voice AI agents, offering models for voice isolation, turn prediction, and interruption detection to improve STT accuracy and natural conversation flow. The RTC SDK focuses on human-to-human communication, providing noise cancellation and accent conversion for contact centers and communication platforms. Additionally, the Voice Translation API offers self-serve, real-time speech-to-speech translation across 61 languages. These models are optimized to run on standard CPUs (x86 and ARM) with low latency, making them suitable for high-concurrency server deployments.
VIVA SDK for Voice AI Agents
- Voice Isolation 2.5: Separates the primary speaker from background noise and secondary voices, reducing word error rate by 46% on average.
- Turn Prediction: Predicts when a speaker is finished directly from raw audio to eliminate awkward pauses in AI interactions.
- Interruption Prediction: Classifies user speech mid-response to tell an AI agent when to stop and listen or keep going.
- Voice Activity Detection (VAD): Robust voice activity detector optimized for on-server CPU deployment.
RTC SDK for Human Calls
- Background Voice Cancellation: Removes background noise and room echo from the outbound audio stream.
- Inbound Noise Cancellation: Strips background noise from inbound mobile and PSTN callers for clear bidirectional audio.
Voice Translation API
- Voice Translation API: Real-time speech-to-speech translation across 61 languages with 96% accuracy on live calls.
- Custom Vocabulary & Dictionary: Allows developers to add domain-specific terms like medication or product names for higher accuracy.
- Built-in BVC: Built-in background voice cancellation handles competing voices and reverberation without preprocessing.
Specifications
- Latency
- 15 ms for Voice Isolation
- Audio Format
- PCM S16LE, 16 kHz mono
- Supported Platforms
- Windows, Mac, Linux, Web (JS/WASM), iOS, Android
- CPU Optimization
- AWS Graviton3 (Arm NEON), Intel Sapphire Rapids (AVX-512)
Plans
- Starter
- Pay as you go: Pay-as-you-go pricing with no monthly fee and $30 in free credits.
- API Access: Access to Voice Isolation and Voice Translation APIs.
- Docs + Community Support: Support provided via documentation and community forums.
- Growth: 400 month
- Included Usage: Includes $400 of usage every month at lower hourly rates.
- Lower Rates: Voice Isolation API at $0.12/hr and Voice Translation API at $3.00/hr.
Pricing
- Starter (Voice Isolation API)
- $0.18 / hr
- Starter (Voice Translation API)
- $4.00 / hr
- Growth (Monthly)
- $400 / month (includes usage)
- VT API Starter (Monthly)
- $249 / mo (45 hrs included)
- VT API Advanced (Monthly)
- $799 / mo (150 hrs included)
Added by AntSpace from public sources (krisp.ai).