Skip to content

Add more Khmer read-aloud prompts for the /voice collector #4

Description

@sengtha

What

Add more short Khmer sentences to the read-aloud prompt set used by the /voice speech-collection page, in src/assets/voicePrompts.ts.

Why

More diverse sentences = better phonetic coverage = a better open Khmer speech-to-text model. This directly helps the community dataset.

Guidelines (from the file's header)

  • Pure Khmer script (write foreign words in Khmer), 3–14 words, one clause.
  • Everyday vocabulary; include numbers, dates, places, and a good spread of Khmer consonants, subscripts (coeng) and vowel signs.
  • No rare proper nouns that trip young readers.

Where to start

  • src/assets/voicePrompts.ts — add lines to the relevant group in GROUPS (ids are generated; just add sentences). A teacher can also add a whole new group.
  • Context: docs/VOICE-COLLECTION.md.

Acceptance

  • New sentences added following the guidelines; npm run typecheck passes.

Perfect for Khmer speakers / teachers — you don't need deep coding experience; comment with sentences if you'd rather not open a PR and we'll add them. 🎤

Metadata

Metadata

Assignees

No one assigned

    Labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions