ASR Evaluation Metrics Voice AI Tools

Tools for measuring and analyzing the accuracy of automatic speech recognition systems through metrics like WER, CER, and DER. Does NOT include ASR models themselves, transcription services, or general audio quality assessment.

There are 26 asr evaluation metrics tools tracked. 6 score above 50 (established tier). The highest-rated is fgnt/meeteval at 63/100 with 149 stars and 5,213 monthly downloads.

Get all 26 projects as JSON

curl "https://pt-edge.onrender.com/api/v1/datasets/quality?domain=voice-ai&subcategory=asr-evaluation-metrics&limit=20"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.

#	Tool	Score	Tier	Stars	Language
1	fgnt/meeteval MeetEval - A meeting transcription evaluation toolkit	63	Established	149	Python
2	readbeyond/aeneas aeneas is a Python/C library and a set of tools to automagically synchronize...	63	Established	2,811	Python
3	analyticsinmotion/werpy 🐍📦 Ultra-fast Python package for calculating and analyzing the Word Error...	63	Established	23	Python
4	kahne/fastwer A PyPI package for fast word/character error rate (WER/CER) calculation	62	Established	70	Python
5	tabahi/bournemouth-forced-aligner Extract phoneme-level timestamps from speeh audio.	59	Established	121	Python
6	wq2012/SimpleDER A lightweight library to compute Diarization Error Rate (DER).	56	Established	62	Python
7	analyticsinmotion/werx 🐍📦 Easy-to-use Python package for lightning-fast Word Error Rate (WER) analysis	49	Emerging	8	Python
8	OpenBMB/UltraEval-Audio Your faithful, impartial partner for audio evaluation — know yourself, know...	42	Emerging	281	Python
9	belambert/asr-evaluation Python module for evaluating ASR hypotheses (e.g. word error rate, word...	42	Emerging	283	Python
10	symblai/speech-recognition-evaluation Evaluate results from ASR/Speech-to-Text quickly	40	Emerging	41	JavaScript
11	hyeonsangjeon/computing-Korean-STT-error-rates STT 한글 문장 인식기 출력 스크립트의 외자 오류율(CER), 단어 오류율(WER)을 계산하는 Python 함수 패키지	33	Emerging	68	Python
12	tts-hub/monotonic_alignment_search Monotonically align text and speech	30	Emerging	4	Python
13	jumon/pywer A simple Python package to calculate word error rate (WER).	30	Emerging	5	Python
14	anilkeshwani/speech-text-alignment Functionality for speech data processing including time alignment, encoding...	25	Experimental	4	Python
15	dangkhoadl/WER-in-cpp Calculates the word error rate between the reference and hypothesis in ASR,...	24	Experimental	2	C++
16	slanglabs-projects/asr-wer-bench Workbench for benchmarking Word Error Rate (WER) of Automatic Speech...	24	Experimental	4	Jupyter Notebook
17	codycollier/wer A word error rate util for golang	23	Experimental	2	Go
18	2017fandrei/ForcedAlignment Graphical utility for forced alignment using aeneas, an interactive audio player	23	Experimental	6	Python
19	narVidhai/Speech-Transcription-Benchmarking Example python scripts to evaluate various ASR methods	21	Experimental	11	Python
20	wq2012/mdeval Python implementation of the NIST md-eval.pl script for evaluating rich...	20	Experimental	1	Python
21	enesyugan/PIER-CodeSwitching-Evaluation Point of Interest Error Rate (PIER) Metric for Code-Switching ASR: A...	19	Experimental	5	Python
22	thibault-roux/metric-evaluator Metric evaluator for Automatic Speech Recognition using the HATS dataset	15	Experimental	6	Python
23	MahtaFetrat/Mana-Forced-Aligner A robust forced alignment tool for low-resource languages using multiple ASR...	15	Experimental	6	Jupyter Notebook
24	oplatek/moosenet-plda Official repo for “MooseNet: A Trainable Metric for Synthesized Speech with...	13	Experimental	6	Python
25	ikegami-yukino/csj-eval For evaluating speech recognition system using the Corpus of Spontaneous...	12	Experimental	3	Python
26	shahruk10/go-sctk Go CLI wrapper around SCTK binaries for word error rate evaluation and error...	12	Experimental	3	Go

Comparisons in this category

aeneas and ForcedAlignment (63 vs 23) werpy and fastwer (63 vs 62) fastwer and werx (62 vs 49) fastwer and pywer (62 vs 30) werpy and werx (63 vs 49) werpy and pywer (63 vs 30) werx and pywer (49 vs 30)