音声
hifi-gan
GAN vocoder that turns mel-spectrograms into high-fidelity speech audio, fast enough for real-time use.
★ 2.4k スターMIT
ガイドを見る →espeak-ngは、espeak-ngが公開しているオープンソースの音声プロジェクトです（GitHub: espeak-ng/espeak-ng）。GitHubスターは6,929、フォークは1,314で、GPL-3.0ライセンスの下で公開されています。GitHubでの最終更新は2026年9月22日です。
eSpeak NG — the compact open-source speech synthesizer for 100+ languages, small enough to embed anywhere.
Pick your language and voice, and decide your route: a packaged install for your OS, or building from source to embed the library.
Clone espeak-ng/espeak-ng from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.
Install eSpeak NG from your platform's package manager or releases, or clone and build the C source following the project documentation.
Synthesise one sentence to a WAV file in your language, listen to it, then adjust voice characteristics to taste.
Supports more than a hundred languages and accents, many unavailable in heavier neural TTS systems.
The program and its data, including many languages, total only a few megabytes — practical for embedded and offline use.
Produces speech output as WAV files and accepts SSML (partially) and HTML input, not just plain text.
Includes different voices whose characteristics can be altered to change how the speech sounds.
Can translate text into phoneme codes with pitch and length, so it can front other synthesis engines such as MBROLA voices.
A portable C codebase used across desktop, mobile (including Android) and embedded contexts.
Get the source from the official GitHub repository for espeak-ng.
git clone https://github.com/espeak-ng/espeak-ng.git
cd espeak-ngInstall the packaged build for your operating system, or clone the repository to build from source per the documentation.
git clone https://github.com/espeak-ng/espeak-ng.git
cd espeak-ng # then build per the project docs — see READMERun the synthesizer on a sentence and write it to a WAV file to verify your language and voice.
espeak-ng "Hello world" -w out.wav # command style per project documentation — see READMEThe espeak-ng community project, through the espeak-ng/espeak-ng repository.
Yes — the project is open source under the GPL-3.0 license as recorded for this listing. See the license terms for any conditions on redistribution.
No keys or network are needed — synthesis is fully local and free under GPL-3.0; note the GPL terms if you redistribute it inside a product.
Neural models sound far more natural but need models, data and compute per language. eSpeak NG sounds robotic by comparison, yet covers 100+ languages in a few megabytes, runs anywhere, and works fully offline — reach for it for coverage and footprint, not studio quality.
Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — synthesise one sentence to a WAV file in your language, listen to it, then adjust voice characteristics to taste.
GAN vocoder that turns mel-spectrograms into high-fidelity speech audio, fast enough for real-time use.
ToucanTTS: fast, controllable text-to-speech for over 7,000 languages, from Stuttgart's IMS.
Free, open source, 100% offline AI text-to-speech for macOS, Windows & Linux.
Double-click any word for definitions, speech, translation, and Anki.
WaveRNN vocoder plus Tacotron TTS in PyTorch: train your own neural speech synthesis on LJSpeech.
Capable AI keyboard for reliable speech transcription.
LibreAIカタログに掲載された音声プロジェクトの中で最もスターの多いもの — espeak-ngが合わなかった場合に検討したい、現実的な候補リストです。
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Robust Speech Recognition via Large-Scale Weak Supervision
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
Port of OpenAI's Whisper model in C/C++
カタログの中でespeak-ngとスター数が近い音声プロジェクト — 前後の棚もそのまま見て回れます。
ODS V3 Pre-Release: Public testing and refinement ahead of the official V3 launch. Turn your PC, Mac, or Linux box into a private AI server.
On-device Speech AI for Apple Silicon
MeloTTS — multilingual text-to-speech for English, Spanish, French, Chinese, Japanese and Korean.
Human-level text-to-speech through style diffusion and adversarial training with speech language models.
VITS — the original end-to-end text-to-speech model: a conditional VAE with adversarial learning, in PyTorch.
Silero Models: pre-trained text-to-speech models made embarrassingly simple