About this skill

Problem

Elementary and middle school English speech practice often lacks quantifiable feedback: students may notice mistakes but not whether they are at the word, sentence, or phoneme level, and applications or teachers may need audio results as structured scores. This skill targets that workflow by sending a reference text and an audio URL or file to a remote scoring service.

How It Works

The core action is a POST request using multipart/form-data, with these key parameters:
- target_text: required reference text for scoring.
- audio_url or audio_file: choose one source, either a remote audio URL or binary audio upload.
- mode: defaults to A for word and sentence scoring; use D for phoneme details or C for discourse-level scoring.

The workflow is to confirm the evaluation need, prepare the standard text and audio source, select the calling method, set the mode, send the request, then parse the total score and detailed dimensions.

Boundaries and Notes

This skill fits speech scoring when a reference text and an accessible audio file or URL are available. audio_url and audio_file must be provided exclusively, not together. Supported formats include mp3/wav/pcm/silk/opus/amrnb/speex; a 16K sample rate and 64kbps mp3 is recommended, with no tag information and no stereo. If ret != 0, inspect msg for the error reason.

Screenshots

Use Cases

Teachers assign reading practice and need to score student mp3 files against the lesson text for total and word/sentence scores.
Education app developers integrate a remote API to convert a reference text and audio URL into structured scores for the UI.
A learning tool needs phoneme-level diagnostics and requests mode D to parse phoneme details for mispronunciation hints.
Course operations process batches of student audio and use ret/msg to separate failures from successful total scores.

Best For

English teachers who need to turn student reading audio and lesson text into total and word/sentence scores.
Education app developers who need to call a remote scoring API and parse structured results in practice workflows.
Learning tool engineers who need to choose A/D/C scoring granularity and handle audio formats and failure responses.
Course operations staff who need to batch-submit student audio, aggregate successful totals, and troubleshoot failures.