Lari Speech to Text: Preserving a Minority Iranian Language with AI

Lari, also known as Larestani, is an Iranian language spoken by approximately 100,000 people in the southern province of Fars, Iran, and in emigrant communities across Kuwait, Qatar, and the UAE. It belongs to the Southwestern Iranian branch, closely related to Persian but retaining archaic features from Middle Persian and even Old Persian. The language uses the Arabic script (Perso-Arabic) which reads right to left, adding a layer of complexity for speech recognition.

Why Accurate Lari Speech to Text Matters

Minority languages like Lari face a double challenge: limited digital representation and lack of commercial NLP tools. Without transcription software, Lari speakers cannot easily convert spoken content into text for subtitles, archives, or learning materials. This is where Speechyou steps in with a dedicated Lari speech to text model that understands the unique phonetic and orthographic features of the language.

Accurate transcription matters for:

• Cultural preservation: Lari poetry, proverbs, and oral histories can be captured verbatim.
• Education: Teachers of Lari can create subtitled videos and interactive transcripts.
• Accessibility: Deaf and hard-of-hearing Lari speakers get subtitles in their mother tongue.
• Media: Lari radio and YouTube channels become searchable and indexable.

Transcription Challenges Unique to Lari

Lari presents several hurdles for automatic speech recognition (ASR):

• Arabic script ambiguities: Short vowels are typically omitted in writing, so the same consonant string can represent multiple words. Speechyou uses a language model trained on Lari text to predict the correct word based on context.
• Dialectal diversity: The four main dialects — Khonji, Evazi, Bastaki, and Kookherdi — differ in vowel quality, consonant inventory, and even basic vocabulary. Speechyou’s acoustic model is trained on recordings from each dialect area.
• Low resource: Public Lari datasets are scarce. Speechyou compensates with transfer learning from Persian and continual user feedback retraining.
• Borrowed sounds: Lari includes Arabic pharyngeals (e.g., /ʕ/) and uvulars that are absent in many languages. Generic ASR systems confuse them; Speechyou’s model has been explicitly tuned for these sounds.

Real-World Use Cases

Oral History Documentation

Lari elders hold a wealth of traditions. Recording and transcribing their stories with Speechyou creates a permanent digital archive. Export subtitles in SRT to pair with video interviews.

YouTube Subtitles in Lari

Content creators can upload Lari-language videos and instantly get Lari subtitle generation with SRT and VTT files. This boosts discoverability and satisfies accessibility guidelines.

Research Interviews

Linguists and anthropologists conducting fieldwork can transcribe hours of Lari interviews in minutes, freeing time for analysis.

Community Radio Automation

Local radio stations broadcasting in Lari can generate automated transcripts for archives or live subtitling via Speechyou’s API.

How Speechyou Helps

Speechyou is the first commercial transcription tool to offer reliable Lari audio to text conversion. It supports:

• Batch processing: Upload multiple files for bulk transcription.
• Dialect selection: Choose the primary dialect for better accuracy.
• Subtitle export: SRT and VTT in Arabic script, right-to-left.
• Unlimited usage with the Solo plan — no per-minute charges.

Conclusion

Lari is a living language with a deep cultural heritage. By providing accurate speech-to-text and subtitling, Speechyou empowers its speakers to document, teach, and share their language in the digital age. Try transcribing your first Lari audio file today and see how AI can preserve minority voices.