How accurate are the diacritics?

Diacritic accuracy is critical in Vietnamese — ma (ghost), mà (but), má (cheek), mả (tomb), mã (horse), and mạ (rice seedling) are six different words. The AI achieves high accuracy on tone marks and vowel quality marks, including stacked diacritics like ỗ, ắ, and ệ. Accuracy is strongest with clear audio and standard pronunciation.

Does it detect Northern vs Southern Vietnamese?

The AI handles both major dialect groups without requiring you to specify one. Regardless of whether the speaker uses Hanoi pronunciation (distinguishing /z/ and /v/) or Saigon pronunciation (merging them), the output follows standard Vietnamese orthography. The spelling is the same — only pronunciation differs.

How does it handle the Đ/đ vs D/d distinction?

Đ/đ and D/d are separate letters in Vietnamese with different pronunciations and meanings. The AI maintains this distinction — producing đi (to go), đường (road, sugar), and đã (already) with the correct stroke-through character, not substituting plain d.

Does it use old-style or new-style tone mark placement?

Vietnamese had a spelling reform affecting where tone marks sit on diphthongs — "hoá" (old style, mark on á) vs "hóa" (new style, mark on ó). The AI generally follows the newer convention endorsed by Vietnamese government publications, placing tone marks on the main vowel of diphthongs.

How does it handle Sino-Vietnamese vocabulary?

Sino-Vietnamese words (Hán Việt) are Chinese-origin vocabulary that form a large part of formal and academic Vietnamese — words like quốc gia (nation), đại học (university), and kinh tế (economy). The AI transcribes these in standard Vietnamese orthography with correct diacritics, like any other Vietnamese word.

Transcribe Vietnamese audio and video to text

Upload any Vietnamese recording and get a precise transcript with every diacritic in place. Vocova's AI distinguishes all six tones and handles Northern, Central, and Southern pronunciation patterns.

Drop your file here or click to browse

.mp3, .wav, .m4a, .aac, .ogg, .flac, .mp4, .mov, .avi, .mkv, .webm·up to 500MB

Vietnamese transcription where every diacritic matters

Vietnamese is the most diacritic-dense Latin-script language in widespread use. Six tones combine with vowel quality marks to produce stacked diacritics like ỗ, ắ, and ệ — and every mark changes meaning. The six words ma, mà, má, mả, mã, mạ each mean something different (ghost, but, cheek, tomb, horse, rice seedling). Regional pronunciation adds another layer: Hanoi distinguishes /z/ from /v/ while Saigon merges them, and Southern speakers often merge final consonants that are distinct in the North. Vocova's AI navigates all of this, producing text where every accent mark and every đ vs d distinction is correct.

How it works

Upload your Vietnamese recording

Drag and drop or select any file containing Vietnamese speech. All major audio and video formats are accepted.

MP3, WAV, M4A, MP4, MOV, MKV, and all other formats
Files up to 500MB supported
No format conversion needed

AI transcribes with full diacritics

The AI processes your audio, identifies all six tones, and outputs Vietnamese text with complete diacritical marks — including stacked tone and vowel marks on the same character.

Full diacritics with correct tone marks on every syllable
Handles Northern, Central, and Southern accents
Speaker diarization for multi-person recordings

Export your transcript

Review your Vietnamese transcript, make any corrections, and export in the format you need. Diacritics are preserved across all formats.

Export as TXT, SRT, VTT, DOCX, or PDF
Timestamps for every segment
Edit directly in the browser before exporting

Features

Complete diacritical accuracy

Vietnamese diacritics encode both tone and vowel quality, and they stack — ỗ carries both a horn (ơ) and a tilde tone mark. Missing or wrong marks change meaning entirely: bàn (table) vs bán (to sell) vs bạn (friend). The AI places every mark correctly across all six tones.

Northern and Southern accent handling

Hanoi and Saigon Vietnamese have significant phonemic differences. Northern speakers distinguish /z/ from /v/ and /ʂ/ from /s/, while Southern speakers often merge these. Final -n/-ng and -t/-c distinctions also collapse in the South. The AI produces correct spelling regardless of which regional pronunciation is in the recording.

Correct Đ/đ and D/d distinction

Vietnamese has two d-letters: Đ/đ (pronounced /ɗ/ in the North, /j/ in the South) and D/d (pronounced /z/ in the North, /j/ in the South). Confusing them changes words — đi (to go) vs di (to move/shift). The AI maintains this critical orthographic distinction.

Label Vietnamese speakers

Multiple speakers are automatically detected and labeled, making conversations, interviews, and multi-host podcasts easy to follow even when speakers have different regional accents.

Handles compound word spacing

Vietnamese is monosyllabic but forms compound words where spacing can be ambiguous — "bàn tay" (hand) is one concept written as two syllables. The AI follows standard Vietnamese orthographic conventions for compound spacing.

Why choose Vocova

Transcribe Vietnamese business communications

Convert Vietnamese meetings, stakeholder calls, and presentations into documented records with every diacritic intact. Works whether your team is based in Hanoi, Ho Chi Minh City, or Da Nang.

Subtitle Vietnamese video content

Generate subtitle files from Vietnamese recordings for YouTube, TikTok, and streaming platforms. Full diacritics in SRT/VTT output mean your subtitles render correctly on every device.

Process Vietnamese media for research

Turn Vietnamese news broadcasts, podcasts, and interviews into searchable text for media monitoring, academic analysis, and content strategy in Vietnam's fast-growing digital market.

Support the Vietnamese diaspora

Transcribe Vietnamese content for overseas communities who need written text for accessibility, family archives, and cultural preservation. Capture spoken heritage in accurate written form.

Who can benefit

Businesses operating in Vietnam

Transcribe Vietnamese meetings, training sessions, and customer interactions into written records. Capture accurate diacritics so documentation is unambiguous and professional.

Vietnamese content creators

Generate subtitles and written content from Vietnamese-language videos and podcasts. Reach wider audiences with properly diacriticized text that reads naturally.

Vietnamese language learners and researchers

Study how tones, diacritics, and regional pronunciation map to written Vietnamese. Build reading fluency by comparing spoken audio to its fully marked written form.

Journalists and translators covering Vietnam

Transcribe Vietnamese sources, interviews, and press events into text for reporting, translation, and fact-checking. Correct diacritics eliminate ambiguity in names and places.