offline-speech

Theme

Download once, bookmark this page, and turn text into speech with no internet. Your text and audio stay on this device.

Preparing step 1: download speech files.

1. Download speech files

Start here: press Download for offline use. This saves the model and voices to disk once; it does not load them into RAM.

Checking saved files…

Offline files are stored in this browser profile. Browser settings or storage pressure can clear them.

2. Load the model

After the download finishes, press Load model. Wait for “Ready”, then go to step 3. Load it again whenever you reopen this page or unload RAM.

Disk · saved offline files: Checking…
RAM · model and runtime: Model unloaded

The page still uses some memory. Loading speech typically needs about 1 GiB of RAM on a desktop; this is an estimate, not a device reading.

Unloading keeps your saved files for quick offline reuse. Deleting removes the downloaded model, voices and libraries and unloads RAM. Only the small page shell remains; offline speech will need another download.

Storage details

Saved-file sizes exclude browser bookkeeping and temporary files.

3. Type text and hear speech

Choose a voice and press Speak. Speak prepares and plays the audio. Recent text-and-voice combinations replay instantly from the beginning without loading or running the model. Save WAV keeps the completed file. Progress is shown below while it is preparing.

Text to speak (up to 20,000 characters)

Save WAV

Preparing offline storage…

Licenses and dependencies

offline-speech’s own code is licensed under GPL-3.0-or-later (GNU General Public License version 3 or later). Third-party components retain their respective licenses.

Speech model: onnx-community/Kokoro-82M-v1.0-ONNX — Apache-2.0. Kokoro generates speech; the bundled model and six voice files use this license.
Text-to-phoneme engine: espeak-ng/espeak-ng (eSpeak NG) — GPL-3.0-or-later. Its code and language data are embedded through phonemizer. The phonemizer JavaScript wrapper separately uses Apache-2.0.

Browser code, model and voice dependencies

Component	Version / revision	License
PRN UI (adapted palette and button geometry)	986d93b681cb49f28b8fcd3069e2b3a04dad1a3f	MIT
Kokoro-82M model and six voices	1939ad2a8e416c0acfeecc08a694d14ef25f2231	Apache-2.0
kokoro-js	1.2.1	Apache-2.0
@huggingface/transformers	3.8.1	Apache-2.0
phonemizer JavaScript wrapper	1.2.1	Apache-2.0
eSpeak NG code and language data	Embedded in phonemizer 1.2.1; engine revision not identified upstream	GPL-3.0-or-later
ONNX Runtime Web (JavaScript and WASM)	1.22.0-dev.20250409-89f8206ba4	MIT
onnxruntime-common (inside Transformers.js)	1.21.0	MIT
@huggingface/jinja (inside Transformers.js)	0.5.3	MIT

Included voices: af_heart, af_bella, af_nicole, am_michael, am_fenrir, bf_emma. Browser bundles are fetched with pinned SHA-256 hashes. offline-speech changes packaging/imports, adds an explicit phoneme-progress callback, and manages hosted voice caching.

The full ONNX Runtime notices below include its upstream dependency inventory across runtime variants. They are a superset, not a claim that every listed native library ships in this browser build. The exact embedded eSpeak NG engine revision is not identified by the published phonemizer bundle.

Development and test dependencies (80 npm packages)

These tools are not shipped to app visitors. Node.js 20 or later supplies build utilities; puppeteer-core and its complete locked dependency graph support browser tests. Tests use a separately installed Chrome/Chromium browser under that distribution’s own licenses.

Component	Version / revision	License
@puppeteer/browsers	2.10.5	Apache-2.0
@tootallnate/quickjs-emscripten	0.23.0	MIT
@types/node	26.6.6	MIT
@types/yauzl	2.10.3	MIT
agent-base	7.1.4	MIT
ansi-regex	5.0.1	MIT
ansi-styles	4.3.0	MIT
ast-types	0.13.4	MIT
b4a	1.9.0	Apache-2.0
bare-buffer	3.7.1	Apache-2.0
bare-events	2.9.2	Apache-2.0
bare-fs	4.8.4	Apache-2.0
bare-path	3.1.2	Apache-2.0
bare-stream	2.13.4	Apache-2.0
bare-url	2.5.6	Apache-2.0
basic-ftp	5.3.1	MIT
buffer-crc32	0.2.13	MIT
chromium-bidi	5.1.0	Apache-2.0
cliui	8.0.1	ISC
color-convert	2.0.1	MIT
color-name	1.1.4	MIT
data-uri-to-buffer	6.0.2	MIT
debug	4.4.3	MIT
degenerator	5.0.1	MIT
devtools-protocol	0.0.1452169	BSD-3-Clause
emoji-regex	8.0.0	MIT
end-of-stream	1.4.5	MIT
escalade	3.2.0	MIT
escodegen	2.1.0	BSD-2-Clause
esprima	4.0.1	BSD-2-Clause
estraverse	5.3.0	BSD-2-Clause
esutils	2.0.3	BSD-2-Clause
events-universal	1.0.1	Apache-2.0
extract-zip	2.0.1	BSD-2-Clause
fast-fifo	1.3.2	MIT
fd-slicer	1.1.0	MIT
get-caller-file	2.0.5	ISC
get-stream	5.2.0	MIT
get-uri	6.0.5	MIT
http-proxy-agent	7.0.2	MIT
https-proxy-agent	7.0.6	MIT
ip-address	10.7.6	MIT
is-fullwidth-code-point	3.0.0	MIT
lru-cache	7.18.3	ISC
mitt	3.0.1	MIT
ms	2.1.3	MIT
netmask	2.1.1	MIT
once	1.4.0	ISC
pac-proxy-agent	7.2.0	MIT
pac-resolver	7.0.1	MIT
pend	1.2.0	MIT
progress	2.0.3	MIT
proxy-agent	6.5.0	MIT
proxy-from-env	1.1.0	MIT
pump	3.0.4	MIT
puppeteer-core	24.10.0	Apache-2.0
require-directory	2.1.1	MIT
semver	7.8.6	ISC
smart-buffer	4.2.0	MIT
socks	2.8.10	MIT
socks-proxy-agent	8.0.5	MIT
source-map	0.6.1	BSD-3-Clause
streamx	2.28.1	MIT
string-width	4.2.3	MIT
strip-ansi	6.0.1	MIT
tar-fs	3.1.3	MIT
tar-stream	3.2.2	MIT
teex	1.0.1	MIT
text-decoder	1.2.7	Apache-2.0
tslib	2.9.1	0BSD
typed-query-selector	2.12.3	MIT
undici-types	8.9.0	MIT
wrap-ansi	7.0.0	MIT
wrappy	1.0.2	ISC
ws	8.22.1	MIT
y18n	5.0.8	ISC
yargs	17.7.3	MIT
yargs-parser	21.1.1	ISC
yauzl	2.10.0	MIT
zod	3.25.76	MIT

Full offline-speech license · Full third-party notices and license texts. Both files are saved with the page for offline reading.

When redistributing this app, preserve notices and meet applicable GPL corresponding-source requirements. Source links identify upstream projects; notices alone do not supply corresponding source for a particular build. A separate HTML page does not by itself determine the license obligations of a containing site. GNU GPL guidance on separate and combined works.