Welsh (Cymraeg) review centers on this evidence: verify initial mutations, digraphs, vowel length where marked, place names, compact particles, and Welsh–English switching. For underspecified cy, CLDR supplies Latn (Latin script) as a likely/default web-locale hint—not an audio attribute or exclusive Welsh orthography. Confirm the actual orthography and script with the project owner before normalizing any text. Orthographic targets include diacritics, word boundaries, proper names, number expressions, and code-switched terms. Project decision: which Welsh mutation, apostrophe, and English-name style the project follows. Before batch editing cy, a reviewer must record which Welsh mutation, apostrophe, and English-name style the project follows and attach a source timestamp to disputed text.
Where this workflow fits
- Welsh interviews and oral histories
- Welsh lectures and training recordings
- Source-linked Welsh research review
Welsh orthography and speech checkpoints
verify initial mutations, digraphs, vowel length where marked, place names, compact particles, and Welsh–English switching. Editorial decision: which Welsh mutation, apostrophe, and English-name style the project follows. Visual inspection should cover record the project decision about which Welsh mutation, apostrophe, and English-name style the project follows.
Script choice and source-linked review for Cymraeg
For underspecified cy, CLDR supplies Latn (Latin script) as a likely/default web-locale hint—not an audio attribute or exclusive Welsh orthography. Confirm the actual orthography and script with the project owner before normalizing any text. For oral-history audio containing mutation, dialect vocabulary, and place names, inspect verify initial mutations, digraphs, vowel length where marked, place names, compact particles, and Welsh–English switching. Visually inspect diacritics, word boundaries, proper names, number expressions, and code-switched terms; record the project decision about which Welsh mutation, apostrophe, and English-name style the project follows. Replay a complete clause around every uncertain name, quantity, interruption, or ending.
Frozen deployment boundary
This page is published only because its language is present in capability snapshot m10-2026-07-20.1. It does not extend that catalog or claim a measured accuracy score.
- The cy hint maps to the reviewed cy provider root and preserves cy as the user's transcript language tag.
- CLDR Latn is a likely/default locale hint, not an audio property; capability flags remain per language and model while editors confirm the project's real script and orthography.
Reviewed provider capability
assemblyai · universal-2
root cy · tiers fast, standard, precision
- Automatic detection
- Reviewed no
- Diarization
- Reviewed no
- Word timestamps
- Reviewed yes
Questions from the workflow
What must a Welsh editor decide before correcting the transcript?
mutated word beginnings, compact function words, apostrophes, and regional forms can only be corrected reliably inside the phrase. For Welsh, verify initial mutations, digraphs, vowel length where marked, place names, compact particles, and Welsh–English switching. Project choice: which Welsh mutation, apostrophe, and English-name style the project follows. Keep unresolved forms beside their source timestamps.
Reviewed sources
- IANA Language Subtag Registry · checked 2026-07-20
- AssemblyAI model and supported-language documentation · checked 2026-07-20
- AssemblyAI pre-recorded audio supported languages · checked 2026-07-20
- Unicode CLDR likely-subtags language and script data · checked 2026-07-20
- Welsh language and orthography editorial reference · checked 2026-07-20
- World Atlas of Language Structures Online · checked 2026-07-20