← Speech to text (English) libraries · Reader board

Parakeet 0.12.0 on Speech to text (English)

130 files: text from 130, empty 0, errors 0, hangs 0, crashes 0. 130 files give text to at least one library.

Note: With huggingface-hub 1.31.0: with 1.32+ (new shared cache layout) and current onnxruntime, the model fails to load ("External data path escapes model directory"), as many other ONNX models do. why

🔴 Fails where another library reads text: none

🟡 Empty where another library reads text: none

⚔️ Compare: Parakeet vs faster-whisper · Parakeet vs Whisper

Changes between versions

Only one version tested so far. The next PyPI release is tested automatically.

Badge

For a README: badge

[![Reader board](https://img.shields.io/endpoint?url=https://prometheus-maker.github.io/reader-board/badges/speech-parakeet.json)](https://prometheus-maker.github.io/reader-board/speech-parakeet.html)

Slowest files

ami-close-mic/AMI_EN2002a_H00_MEE073_0030336_0030633.wav	4.15 s, 2552 MB
librispeech-clean/5683-32866-0019.wav	2.475 s, 2740 MB
librispeech-other/5484-24317-0027.wav	2.375 s, 2753 MB
librispeech-clean/1089-134691-0009.wav	2.352 s, 2675 MB
fleurs/11318597463051480163.wav	2.169 s, 2669 MB
librispeech-clean/1221-135767-0003.wav	2.122 s, 2678 MB
earnings-calls/457_12_45.wav	2.117 s, 2652 MB
earnings-calls/735_10_43.wav	2.025 s, 2667 MB
fleurs/8398257681626621326.wav	1.943 s, 2669 MB
earnings-calls/28_8_25.wav	1.867 s, 2648 MB

Recipe: board/targets.py. Run: 2026-11-06T19:46:41Z, Python 3.13.18, Linux x86_64, 4 CPUs (log). Installed: PyYAML==6.0.3, RapidFuzz==3.14.6, anyio==4.16.0, certifi==2026.10.23, click==8.5.1, filelock==4.2.9, flatbuffers==25.12.19, fsspec==2026.9.0, h11==0.16.0, hf-xet==1.8.0, httpcore==1.0.9, httpx==0.28.1, huggingface_hub==1.31.0, idna==3.20, indic_numtowords==1.1.0, more-itertools==11.2.1, numpy==2.5.4, onnx-asr==0.12.0, onnxruntime==1.31.1, packaging==26.3, protobuf==7.36.2, regex==2026.9.29, text2num==3.1.0, tqdm==4.70.1, typing_extensions==4.16.0, whisper_normalizer==1.0.0