dsh-tingdong: A DSH plugin that turns English audio and documents into listening lessons in DeepSeek Harness
aicivilization/dsh-tingdong
A DSH plugin that turns English audiobook MP3s or documents into sentence-by-sentence listening lessons with local speech recognition, AI translation and spaced repetition.
dsh-tingdong is an English listening plugin for DeepSeek Harness that turns MP3 audiobooks or English documents into lessons you can study sentence by sentence. It targets the gap left when nobody explains the audio for you: every English sentence and its timing is recognized locally without uploading audio, then the model you selected in DSH proofreads, translates, extracts vocabulary and writes hints about linking and weak forms. Daily study, intensive listening, spaced repetition and passive listening modes come with subtitle and Chinese dubbing export, while lessons and progress stay on your machine.
How to Install
dsh plugin --profile web add dsh-tingdong- Category
- Tools & Capabilities
- Platform
- DSH Plugin
- Author
- aicivilization
- Distribution
- Plugin
dsh-tingdong Key Features
dsh-tingdong Repo Summary
What Does It Do?
Tingdong is a DSH plugin for DeepSeek Harness that turns English audiobooks (MP3) or English documents into listening lessons you can study sentence by sentence. It solves the problem of having no one to explain English audio: it recognizes every sentence locally with start and end times, then uses the model you selected in DSH to proofread, translate, pick out vocabulary and write listening hints, and pairs this with spaced repetition until you understand by ear. Lessons and progress stay on your machine, audio is never uploaded, and it also works outside DSH.
Core Features
- Two import paths: accepts MP3 audio plus PDF, Word, EPUB, HTML, TXT and Markdown documents, and shows a confirmation card with word count, estimated lesson count, narration length and token usage before processing.
- Local sentence recognition: uses the same speech model as DSH voice input to detect each English sentence and its timing, with no audio upload and no extra download if you already use DSH voice input.
- AI sentence explanations: corrects recognition errors, translates into conversational Chinese, extracts vocabulary with phonetics, and writes listening hints for linking, weak forms and elision, plus a whole-lesson summary and glossary.
- Multiple practice modes: daily study, intensive listening (listen first, then check), spaced-repetition review with self-rating, meaning choice and dictation, passive listening, continuous playback and vocabulary review.
- Export and dubbing: exports bilingual SRT subtitles, LRC lyrics, recording scripts and CSV, and can replace English with Chinese dubbing as MP3 or lossless FLAC.
- In-chat tools: registers tools for importing audio, importing text, listing lessons and reading sentences, so you can simply ask the AI in a DSH conversation to organize a PDF and import it.
How to Use This Plugin?
After enabling it in DSH, open the right sidebar in the chat page and choose the Tingdong English listening panel. Pick import audio or import document on the home page, confirm the word count and time estimate, and processing runs in the background so you can start studying the first finished lesson. During practice, space replays, Enter advances, 1/2/3 rate your comprehension, arrow keys change sentences, S adjusts speed and L loops one sentence; the card menu lets you favorite a sentence or fix a misrecognized one. The lesson header switches between original audio, silence-trimmed audio, system speech and offline narration, and the export tab provides subtitles and Chinese dubbing downloads. Progress is stored locally, so you can continue on another device or in a mobile browser.
How to Troubleshoot This Plugin?
Scanned PDFs contain page images and yield no text, so convert them with OCR or AI first. Files over 100 MB or text beyond roughly 60,000 English words should be split into several imports. MP3 decoding depends on ffmpeg; when it is missing, the Tingdong section in settings reports it, and installing it per the official instructions restores recognition. The Edge speech component needs a network connection and a local Python environment, so switch to offline narration if dubbing fails.
This page is an independent rewrite of the plugin's official README — for authoritative documentation and the latest changes, refer to the source: aicivilization/dsh-tingdong. The plugin is third-party code that runs on your machine once installed; inclusion does not imply endorsement — please review the source before installing.