The unified platform for real-time Voice AI and AI agents.
Build real-time voice applications with one integrated stack. Decibri brings together audio capture, processing, streaming, orchestration, AI models, and integration across Python, Node.js, Rust and browsers.
From microphone to model to speaker, Decibri provides the foundation for building intelligent, voice-enabled applications.
⌬ decibri
Core Voice AI Audio Engine | Rust, Python, Node.js
Cross-platform audio infrastructure for real-time Voice AI. Capture, process, stream and play audio across Python, Node.js, Rust and browsers, with built-in voice activity detection (VAD), acoustic echo cancellation (AEC) and audio conditioning.
Cross-Platform Audio Command Line Tools
Capture, play and manage audio devices directly from your terminal. A lightweight, cross-platform CLI built on Decibri, with no runtime dependencies.
Microphone and Speech Tools for AI Agents
Give MCP-compatible AI agents the ability to listen. Integrates microphone capture and speech-to-text capabilities into agent workflows using the Model Context Protocol (MCP).
Explore the core capabilities powering real-time audio, speech processing and AI agent integration.
| Capability | What it enables |
|---|---|
| ✔ Audio capture and playback | Real-time microphone input, speaker output and audio streaming |
| ✔ Multichannel audio | Capture and process mono, stereo and multichannel audio |
| ✔ Voice activity detection (VAD) | Detect speech using Silero or energy-based VAD |
| ✔ Audio Conditioning Engine (ACE) | Condition microphone audio with noise suppression, gain control and filtering |
| ✔ Acoustic echo cancellation (AEC) | Reduce playback echo in two-way audio |
| ✔ Automatic audio decoding | Detect and decode WAV, AIFF and FLAC |
| ✔ Audio File Processor (AFP) | Read, write, condition and analyse audio files |
| ✔ Audio resampling | Convert between sample rates |
| ✔ Cross-platform development | Windows, macOS, Linux and browsers, using Python, Node.js and Rust |
