OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Kaldi Active Grammar

Kaldi Active Grammar is a Python package that combines Kaldi speech recognition with grammars that can be activated or deactivated dynamically for each utterance. It supports context-sensitive voice command and control, including use as a Dragonfly backend.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Kaldi Active Grammar — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a Kaldi-based speech interface needs to switch available command grammars per utterance rather than keep every grammar active during decoding.

Who it is for

  • Developers building voice control for desktop applications.
  • Dragonfly users who need grammars selected for the active context.
  • Builders of voice-coding and other speech interfaces with changing command vocabularies.

Documented uses

  • Enable only the command grammars relevant to an application's current context.
  • Connect recognized voice commands to actions through Dragonfly grammars.
  • Use Kaldi for plain dictation through the documented dictation interface.
Before you choose

Requirements & limits

The project is open source under the AGPL-3.0 license. Its README lists Windows, Linux and macOS support and distributes platform-specific wheels.

  • Requires 64-bit Python 3.6 or later.
  • Requires a compatible, modified Kaldi left-biphone nnet3 chain model.
  • The README estimates about 1 GB or more of disk space and RAM, depending on model and grammar complexity.
  • The Python package includes native Kaldi decoding binaries; source distributions are intentionally unsupported.
  • A general English model is available through project releases; standard Kaldi models require conversion, which the README says is not fully implemented.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Kaldi Active Grammar for direct use

Install the package and prepare a compatible model for use through its lower-level interface.

Before you start

  • 64-bit Python 3.6 or newer
  • About 1 GB or more of disk space and 1 GB or more of RAM, depending on model and grammar complexity
  • A compatible Kaldi left-biphone nnet3 chain model with the required modifications

Steps

  1. Download a compatible generic English Kaldi nnet3 chain model from the project releases and unzip it.

  2. Install the package and its included Kaldi binaries with the command below.

    Documented command
    pip install kaldi-active-grammar

  3. Pass the unzipped model directory path to the kaldi-active-grammar constructor.

Expected outcome

The package and compatible model are prepared for direct use through the lower-level interface.

Guide limitations

  • Only 64-bit Python 3.6 or newer is supported.
  • Only Kaldi left-biphone nnet3 chain models with specific modifications are supported; standard Kaldi models require conversion, which is not fully implemented.
  • Resource requirements vary with model and grammar complexity.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
daanzu/kaldi-active-grammar · Official homepage title · retained in its source language
How we review sourcesSuggest a correction