Decibri
Verified
Audio AI Tools
Decibri is a cross-platform audio stack for real-time Voice AI and agents, providing low-latency capture, conditioning, VAD, and PCM streaming in Python, Node.js, and Rust.
Promoted
AI Checker PlusAI detector that highlights which passages AI wrote. Free check, no sign-up. Plagiarism too.
RewindType any X/Twitter handle. Get their year back as a narrated video story. Then ask it anything.
KurodyneSell the tools from your AI side projects. Listing is free.
Pixmax.aiPixmax.ai is an all-in-one AI workspace for videos, visual stories, and marketing content.
SEOify - Improve your SEO & rankingsBuild links to rank in Google. Improve your DR with links you can start building now.Advertise here330,000+ people visit Dang.ai every month.
What does Decibri do?
Decibri is a unified technology stack for building real-time Voice AI applications, intelligent agents, and interactive devices. It provides the core audio infrastructure you need: direct microphone capture, speaker playback, audio conditioning, speech activity detection, and streaming audio for downstream speech or agent services. Instead of stitching together separate OS tools and language-specific audio layers, Decibri gives you a consistent foundation across platforms and languages.
Key features include low-latency audio I/O with direct OS audio access, and predictable chunking (100ms audio chunks by default). Decibri also includes an Audio Conditioning Engine (ACE) for improving audio before it reaches speech models, including voice activity detection (VAD), noise reduction, acoustic echo cancellation (AEC), automatic gain control, multichannel handling, and resampling. You can also process recorded audio through the same conditioning capabilities.
Decibri is designed to fit your speech stack choices. It outputs raw PCM (16-bit signed integer PCM by default, little-endian) so you can stream audio to local engines or cloud speech-to-text providers. You can build fully offline pipelines or hybrid deployments—while keeping control over models, providers, and configurations. It ships with a bundled Silero VAD v6.2 model that runs locally via ONNX Runtime.
Decibri supports Python, Node.js, and Rust with pre-built binaries and a CLI, including browser support for the Node.js package via conditional AudioWorklet exports. Install is one command (e.g., pip install decibri), and it follows a freemium approach with freely available components plus commercial offerings.
What audio format does Decibri output?
Decibri outputs 16-bit signed integer PCM (little-endian by default). A 32-bit float output format is also available for workflows that require it.
How does Decibri capture audio from my microphone?
Decibri accesses the operating system’s audio layer directly through its Audio Host—WASAPI on Windows, CoreAudio on macOS, and ALSA on Linux—so you avoid external audio tools and intermediate encoding steps.
Do I need build tools to install Decibri for Python or Node.js?
No. Pre-built binaries are published for supported platforms, so pip and npm can install without compilers, node-gyp, or post-install downloads. The Rust crate is compiled with cargo.
Does Decibri support running in the browser?
Yes for Node.js and browser use. When bundled for the browser (e.g., webpack or Vite), conditional exports provide the browser implementation (AudioWorklet), so you don’t need a separate package.
Can I select a specific microphone device?
Yes. Use Microphone.devices() to list available inputs, then pass a device index or a name substring when creating your Microphone instance.
Is voice activity detection available offline?
Yes. Decibri includes a bundled Silero VAD v6.2 model that runs locally in Rust via ONNX Runtime, producing speech/silence events without requiring a cloud API.
This listing is claimed
Own Decibri? Sign in with the email used to claim it to manage this listing, or upgrade to Pro for featured-first placement and a newsletter feature.