Kuakua Dictation Assistant

The problem

In after-school, tutoring, and home settings, dictation practice often stalls on repetitive work: a teacher or parent reports words one by one, formats the sheet by hand, and records the listening pass separately. Chinese dictation is especially sensitive because the sheet should show pinyin prompts rather than the target hanzi; English materials also need to separate words, phrases, and sentences, and the audio becomes incomplete without translations or simple example sentences.

How the skill works

The skill turns a dictation request into a concrete pipeline. It first extracts the student name, lesson, and word or sentence list, then routes the content by shape. Chinese terms use the Chinese dictation mode, producing pinyin with Unicode tone marks and writing lines. English words use a four-column word layout, while English sentences use a two-column sentence layout with Chinese prompts. After the PDF is generated, the skill also creates an MP3: Chinese audio reads each term plus a simple sentence twice, and English audio follows the word or sentence pattern with pauses.

The workflow enforces several constraints: Chinese pinyin should use tone-marked output such as hóng shuǐ rather than hong shui; English entries require Chinese translations; Chinese terms require natural example sentences; TTS uses the Volcano Engine Doubao v3 unidirectional endpoint, and audio is decoded only when the response has code == 0 and data.

Fit and caveats

It fits cases where a word or sentence list already exists and the goal is to produce printable worksheets plus matching dictation audio quickly. The documented setup assumes a local Windows environment, including paths such as D:\wtb and C:\Windows\Fonts\msyh.ttc, and environment-based TTS credentials. Do not hard-code secrets; occupied files may get _v2 or _v3 suffixes.