A GGUF quantized version of NVIDIA's Parakeet TDT 0.6B v2 English automatic speech recognition model. It supports local inference using tools such as transcribe.cpp and is suitable for on-device or offline speech-to-text applications. The model is based on Conformer and RNNT architectures and is released under a permissive license.
In the Voice, TTS & speech space, Parakeet Tdt 0.6b takes a focused approach. It focuses on running high-quality speech-to-text transcription locally with reduced memory requirements. It is built as an open-source project for developers and researchers. Parakeet Tdt 0.6b is open source under the MIT license. Parakeet Tdt 0.6b is available on the command line and API.
It is developed by handy-computer, and the product first shipped in 2026. Development happens publicly on GitHub with 1.4k stars and 419 commits in the last 90 days. It operates in a well-populated space: PulseGate tracks 17 similar tools.
Latest indexed changes and source events
handy-computer/parakeet-tdt-0.6b-v2-gguf verified by the PulseGate indexer
