Codeep Voice is a free WordPress text-to-speech plugin for English posts. Your computer generates the audio inside your WordPress admin browser and saves an MP3 in your Media Library. Readers can play or download the file. No TTS account or API key is needed. Keep the admin tab open while it records.
Try eight English Kokoro voices with your own text and download an MP3 on Hugging Face. No WordPress installation needed. The first run downloads the model.
The same sentence, read twice: once by this plugin, and once by the speech voice built into your browser, through the Web Speech API.
Kokoro-82M, the voice called Heart
speechSynthesis, your browser's own
Both clips are levelled to the same loudness, because the louder recording wins on the ear no matter how it sounds.
See and hear Codeep Voice, then try one short English post.
YouTube loads only when you press play. Watch on YouTube.
In WordPress, open Plugins → Add New, search for Codeep Voice by sdsoftware, then install and activate it.
Open Codeep Voice in your admin menu and press Prepare the voice. The screen shows the download size before you start. The voice stays cached in this browser.
Choose a short English post and press Voice it. Keep the tab open until it finishes. The MP3 appears in your Media Library, and a player appears on the post.
Open the published post, listen to it and try the download button. Check names and abbreviations before sharing the audio.
Use an up-to-date browser in your WordPress admin. Graphics acceleration helps, but is not required. If recording is slow, try Amy, the free voice designed for computers without a graphics card. Your readers can listen on phones.
The nine free voices read English. WordPress 6.0 or later and PHP 7.4 or later are required. No account or API key is needed.
Voice files download from Hugging Face before the first recording. Download size and recording time depend on the voice and your computer. Keep the admin tab open while it works.
Check recording speed on your computer before installing, or read the detailed setup guide.
One row above your text. Under 12 KB, it takes its colours from your theme, and the arrow on the right hands the reader the MP3.
It lands in your media library like any upload. Download it, attach it to a podcast feed, hand it to someone. It is yours.
Heart, Bella, Nicole and Sarah; Michael and Adam; Emma and George. Plus Amy, an American voice that is fast on a computer with no graphics card. Set one for the whole site or pick a different one per post.
Under 12 KB. They download the MP3 and nothing else. No model, no engine, no waiting.
Nothing to sign up for and nothing to run out of. Voice as many posts as you like, as often as you like.
The screen inside your WordPress. It reads your hardware, tells you what the download will cost you once, and lists your posts with the ones that already have audio.
The speech is computed by the machine that is already in front of you, inside your own admin screen. We run no server for it, so there is no cost per use on our side to pass on to you, and nothing to count.
| Where the voice is made | in your browser |
| Where your post text goes | nowhere |
| API key | none |
| Account | none |
| Cost per post | nothing |
| Limit | none |
| Outside calls after the first download | none at all |
The last row is the one that matters most, and it is the reason for all the rest. Once the voice model is in your browser, this plugin talks to nothing and nobody. There is no service behind it that could change its price, change its terms, or stop existing.
Every browser can speak text aloud through the Web Speech API, free of charge. You heard it at the top of this page. If speaking aloud is all you need, you do not need this plugin, and we would rather say so.
What that API cannot do is give you a file. The specification defines speech as something the browser performs, not something it produces: there is no output to capture, so there is no MP3 at the end. That rules out a download for your reader, a podcast feed, a CDN, and anything else that begins with having the audio as a file.
That gap is the whole reason this plugin exists. It runs a real model and writes a real file.
Because your computer does the work, the speed is your computer's speed. A machine with graphics acceleration is in a different league from one without, and we are not going to print a number here that turns out to be a lie on your hardware.
So measure it instead. The page below downloads the same voice this plugin uses and times one sentence on your machine. Nothing is installed and nothing is sent anywhere.
Want the setup and technical details? Kokoro in WordPress, running in your browser includes the workflow and original audio test files.
Hear a full 1,000-word article recorded in WordPress, with the test text, computer details and the plugin's completion time.
The voice model itself is not bundled with the plugin. It is fetched from Hugging Face the first time you press Prepare the voice, and then cached in your browser. That transfer is one-way: the model comes in, and nothing about you or your post text goes out. This plugin has no analytics and makes no call home of any kind.
Codeep Voice is available on WordPress.org. In your WordPress admin, open Plugins → Add New, search for Codeep Voice, then install and activate it. No account or API key is needed.