jiwer
Evaluate your speech-to-text system with similarity measures such as word error rate (WER)
Decision gist · record as of 2026-08-14
Yes, for ASR evaluation workflows. The package is stable, has no vulnerabilities, and offers the standard metrics researchers need. The 421-day gap since last release is a minor concern for maintenance, but the code is straightforward and the dependency on RapidFuzz is solid. Install if you need to score speech recognition output; skip if you need active development or additional metrics beyond the five supported.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python >= 3.8
- Low install friction; pure Python wheel with only two runtime dependencies (click, rapidfuzz).
- Package is aging—last release was 421 days ago—but supports current Python versions (3.8+) and has no known vulnerabilities.
License · maintenance · safety
Apache-2.0 (permissive) — Apache-2.0 permissive license allows commercial and private use with minimal restrictions; suitable for most projects.
last release 2025-06-19 (421 days)
0 known vulnerabilities (OSV.dev, 2026-08-14) · 4,676,271 downloads/mo, #2,257 on PyPI
Alternatives
Verify before relying
pip install jiwer
from jiwer import wer
reference = "hello world"
hypothesis = "hello duck"
error = wer(reference, hypothesis)- Whether the 421-day gap since last release reflects active maintenance or dormancy
- Performance characteristics compared to other ASR evaluation libraries
- Whether all five error metrics (WER, MER, WIL, WIP, CER) are equally mature
What it is and what it does
JiWER is a Python package for evaluating automatic speech recognition (ASR) systems by computing standard error metrics. It calculates word error rate (WER), match error rate (MER), word information lost (WIL), word information preserved (WIP), and character error rate (CER) by computing the minimum-edit distance between reference transcripts and ASR hypothesis outputs. The package delegates the computationally expensive edit-distance calculation to RapidFuzz, which uses C++ under the hood, making it faster than pure Python implementations.
The package is designed for researchers and engineers validating ASR model quality. It handles edge cases like empty references (treating them as correct silence predictions) and supports batch processing of multiple reference-hypothesis pairs. With low install friction, no security vulnerabilities, and a permissive Apache 2.0 license, it integrates easily into evaluation pipelines.
Use it for
- Compute WER to benchmark an ASR model against a test set of reference transcriptions
- Evaluate whether a speech model hallucinates on silent audio by testing empty reference-hypothesis pairs
- Compare multiple ASR systems using MER, WIL, and WIP metrics alongside WER
- Integrate ASR evaluation into a CI/CD pipeline to track model quality across releases
- Analyze character-level errors (CER) to diagnose spelling or phonetic issues in ASR output
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, for ASR evaluation workflows.
The package is stable, has no vulnerabilities, and offers the standard metrics researchers need. The 421-day gap since last release is a minor concern for maintenance, but the code is straightforward and the dependency on RapidFuzz is solid. Install if you need to score speech recognition output; skip if you need active development or additional metrics beyond the five supported.
Install
jiwer on PyPI
Before you install
Low install friction; pure Python wheel with only two runtime dependencies (click, rapidfuzz). Package is aging—last release was 421 days ago—but supports current Python versions (3.8+) and has no known vulnerabilities.
Requires Python >= 3.8
License in practice
Apache-2.0 permissive license allows commercial and private use with minimal restrictions; suitable for most projects.
Quickstart
pip install jiwer
from jiwer import wer
reference = "hello world"
hypothesis = "hello duck"
error = wer(reference, hypothesis)Verify before relying
- Whether the 421-day gap since last release reflects active maintenance or dormancy
- Performance characteristics compared to other ASR evaluation libraries
- Whether all five error metrics (WER, MER, WIL, WIP, CER) are equally mature
Package facts
| License | Apache-2.0 permissive |
| Python support | Supports the current Python release >=3.8 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 2 packagesclickrapidfuzz |
| Maintenance | Aging 421 days since the last release |
| First released | |
| Downloads | 4,676,271 / month, #2,257 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
Evidence: jiwer-4.0.0-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “word error rate calculation”
- jiwerJiWER computes speech recognition evaluation metrics (WER, MER, WIL,…
- nlptuttiMeasures Korean speech-to-text accuracy using character error rate…
- kaldialignComputes edit distance, alignment, and word error rate (WER) between…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also kaldialign · nlptutti · texterrors · RapidFuzz · jarowinkler · whisper-timestamped · thefuzz · nemo-toolkit · whisperx · onnx-asr