Kaldi

by Kaldi community

The classic C++ HMM/DNN speech recognition toolkit.

TL;DR

The classic C++ HMM/DNN speech recognition toolkit.

Best for telephony, low-latency streaming, fine-grained WFST decoders. Pricing: free.

Category
Open source
License
—
Stars
—
Last push
—
Pricing
free
Platforms
Linux, macOS

What it is

Kaldi has powered most production ASR for a decade. Still the reference for WFST decoding + speaker pipelines. Apache-2.0.

Best for: Telephony, low-latency streaming, fine-grained WFST decoders.
Watch out for: Older paradigm; most new research has moved to end-to-end frameworks.

Install / use

git clone https://github.com/kaldi-asr/kaldi

Features

Speaker diarizationYes
Word-level timestampsYes
Streaming / real-timeYes
Languages supported40
HIPAA eligibleNo

Kaldi vs Whipscribe

FeatureKaldiWhipscribe
CategoryOpen sourceTranscription APIs
Pricingfreefree beta
Speaker diarizationYesYes
Word timestampsYesYes
StreamingYesNo
Languages4099
PlatformsLinux, macOSWeb, API, MCP

Alternatives to Kaldi

OpenAI Whisper
OpenAI

The reference open-source multilingual ASR model from OpenAI.

OSS · MIT ★ 98.1k
whisper.cpp
Georgi Gerganov

C/C++ port of Whisper — runs on anything, from a Raspberry Pi to Apple Silicon.

OSS · MIT ★ 48.8k
faster-whisper
SYSTRAN

4× faster than reference Whisper using CTranslate2 — production sweet spot.

OSS · MIT ★ 22.3k

Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.