SenseAudio Language Tutor

Create interactive language-learning audio with official SenseAudio TTS endpoints and parameters.

What This Skill Does

• - Generate pronunciation examples in supported voices
• Create bilingual vocabulary and sentence practice audio
• Produce slowed-speed listening drills for learners
• Build short dialogue exercises with repetition pauses
• Export lesson audio files and companion study notes

Credential and Dependency Rules

• - Read the API key from SENSEAUDIO_API_KEY.
• Send auth only as Authorization: Bearer <API_KEY>.
• Do not place API keys in query parameters, logs, or saved examples.
• If Python helpers are used, this skill expects python3, requests, and pydub.
• INLINECODE5 may also require a local audio backend such as ffmpeg; if unavailable, prefer writing individual audio files instead of merging them.

Official TTS Constraints

Use the official SenseAudio TTS rules summarized below:

• - HTTP endpoint: INLINECODE7
• Model: INLINECODE8
• Max text length: 10000 characters
• INLINECODE10 is required
• INLINECODE11 range: INLINECODE12
• Optional audio format values: mp3, wav, pcm, INLINECODE16
• Optional sample rates: 8000, 16000, 22050, 24000, 32000, INLINECODE22
• Optional MP3 bitrates: 32000, 64000, 128000, INLINECODE26
• Optional channels: 1 or INLINECODE28

Recommended Workflow

1. 1. Prepare lesson content:

• - Split vocabulary, example sentences, and dialogues into short chunks.
• Keep each API call comfortably below the 10000 character limit.

1. 2. Build minimal TTS requests:

• - Send model, text, stream, and voice_setting.voice_id.
• Add speed, pitch, vol, and audio_setting only when needed.

1. 3. Decode and save audio safely:

• - HTTP responses return hex-encoded audio in data.audio; decode before saving.
• Keep filenames deterministic and avoid exposing secrets in paths or logs.

1. 4. Compose lessons carefully:

• - If pydub and an audio backend are available, merge clips and insert silence.
• Otherwise, emit per-word or per-sentence clips and a manifest/Markdown study guide.

1. 5. Handle failures and traceability:

• - Check HTTP status and provider error payloads before decoding audio.
• Record trace_id only for troubleshooting and avoid showing it unless needed.

Minimal Helper

CODEBLOCK0

Patterns

Vocabulary Drill

• - Generate one clip for the target word
• Generate one clip for an example sentence
• Optionally generate a slower clip at INLINECODE41
• Save clips separately or merge with pauses

Bilingual Lesson

• - Alternate source phrase and translated phrase
• Use short pauses (1000-2000ms) between clips
• Consider different voice_id values for source and translation when helpful

Dialogue Practice

• - Create one clip per line of dialogue
• Insert repetition pauses after each line
• Prefer shorter turns for easier debugging and regeneration

Output Options

• - Individual MP3 clips for words, sentences, or dialogue turns
• Merged lesson audio if local audio tooling is available
• Markdown study guide with transcript, translation, and file manifest

Safety Notes

• - Do not hardcode credentials.
• Do not claim unsupported language-selection parameters for TTS unless the official docs add them.
• Avoid assuming raw bytes can be passed directly to pydub.AudioSegment; decode and load through a supported container format.