DLE Listen PRO

DLE Listen PRO turns the articles of a DataLife Engine site into audio with neural voices. A modern player appears under every post: a natural voice from Yandex, OpenAI, ElevenLabs, Google or Azure, a waveform, playback speed, highlighting of the paragraph being read and a mini player that stays on screen while you scroll. Audio is created once and stored on your server as mp3 — visitors start listening instantly, and the articles also go out as a podcast. Supports DLE 21.0.

Highlights

• Five voice services and a free mode. Yandex SpeechKit, OpenAI, ElevenLabs, Google and Azure with your own key — or the browser voice with no keys and no costs.
• A player people want to press. A waveform instead of a plain bar, speed, ±15 seconds, resume from where you stopped, paragraph highlighting and a mini player; accent colours, corner radius and dark theme are set in the admin panel with a live preview.
• Costs under control. Voice-over on publish, on the first click or manually; the archive comes with a cost estimate before you start; daily and monthly character limits and an email when a limit is reached.
• Reads it right. A pronunciation and stress dictionary; dates, numbers, units and abbreviations spoken as words; spoilers, quotes, tables and any blocks you choose are skipped.
• New audiences. A podcast feed for Yandex Music and Apple Podcasts, a “Listen” page, playlists and “Listen” buttons right in the post list.
• For search and analytics. AudioObject, og:audio and speakable markup, statistics of plays and completions, access by user group and audio banners.

How it works

The module places the player above the full story by itself — or wherever the template has the {listen} tag. The post text is cleaned of everything unnecessary, normalised and sent to the chosen service in chunks; the finished chunks are joined into a single mp3 with paragraph marks and saved to uploads/listen_pro. Voice-over runs in the background — on cron or on site visits — so pages never wait for it. The service is not called again until the post text changes.

• Neural voice — an mp3 on your server, the same sound on every device, download and podcast.
• Browser voice — free and without files: it reads with whatever the visitor's device can do; the module hides the player if the device has no good voice.
• The mode can be chosen for the whole site or separately for each category.

Features

Player on the site

• Button, title, time, a waveform you can tap to seek, speed 0.75–2×, ±15 seconds, mp3 download.
• Highlighting of the paragraph being read and scrolling to it; Alt+click on a paragraph — listen from that point.
• Resume from where the visitor stopped; a “Listened” mark.
• A mini player at the bottom or top of the screen when the main one scrolls out of view; control from the phone lock screen.
• After the end — the next article from the same category with a countdown and a “Cancel” button.
• Light and dark theme (picks up the site theme by itself), accent colours, background, corner radius and ready-made colour sets; keyboard and screen reader support.

Voice services

• Yandex SpeechKit, OpenAI (gpt-4o-mini-tts, tts-1, tts-1-hd), ElevenLabs, Google Cloud TTS, Azure Speech; voices and models are listed straight from the service.
• A separate voice for each site language, a reading style for models that support it, mp3 quality.
• Proxy and a custom API URL for services that are unavailable from the server's country; a “Listen” button to check the key and the voice.
• A key error or an exhausted quota pauses the queue instead of wasting attempts.

Text

• What to read: title, short and full story, additional fields, an intro and an outro in your own words.
• What to skip: quotes, tables, spoilers, [hide], blocks by CSS selector, fragments by regular expression, your own [nolisten] tag.
• Reading of dates, money, units of measurement and abbreviations for Russian, Ukrainian, English and Spanish.
• A pronunciation dictionary with stress marks, a “how it will be read” check on any post, CSV import and export.

Queue, archive and costs

• Voice-over on publish and edit, on the first “Listen” click, or manually only.
• Archive: choose categories, dates and language, count characters, cost and hours of audio before starting, queue in portions, and a “Stop” button.
• Daily and monthly character limits, an email to the administrator, a work log and spending by day.
• A post list with voice-over status: regenerate, delete, listen, add to the queue by ID.

Pages, podcast and search

• A “Listen” page with all audio versions and a “Listen to all” button.
• An RSS podcast feed with iTunes tags: cover, author, category — one for the whole site and one for each site category.
• Template tags: a “Listen” button and duration in short stories, “has audio / no audio” blocks, a playlist with a choice of categories.
• AudioObject, og:audio and speakable markup on the post page.

Statistics, access and revenue

• Plays, completions at 25 / 50 / 75 / 100 %, minutes listened, phones and browser voice — without cookies.
• Access by user group: who can listen, who can download, who gets no ads; a text and a link for closed groups.
• Audio banners before the article: your own mp3 or a text voiced by the service; categories, dates, frequency, erid labelling, impressions and clicks.

Compatibility

• DLE Multi-Language: a separate voice-over for each translation and a separate voice for each language.
• Posts added by parsers and other modules (such as DLE Parser PRO and Rewrite PRO) are voiced automatically: the module picks up new and edited posts when the queue runs.
• An “Audio” block in the post editor: status, plays and a link to manage it.
• DLE 15.1 – 21.0, PHP 7.4 – 8.5; admin panel in Russian, English, Ukrainian and Spanish.