arabic-transcript-cleanup

Topics in the source

• Arabic Transcript Cleanup
• 1. Get a good transcript first (if the user doesn't have one)
• 2. Typical errors and fixes
• 3. Filler words by dialect (mark them, don't silently delete)
• 4. Output
• Rules

Speech-to-text on Arabic dialect makes predictable mistakes. Fix them before captions, editing, or repurposing. The general filler-cutting method is in filler-removal / transcript-polisher; this skill handles the Arabic part. Always force --language ar. Auto-detect sometimes picks Persian/Urdu for short clips. If editor-apprentice is installed, its transcribe.py already produces word-level transcripts per speech island and flags islands where speech was heard but no words came out. Clean that output here (sections 2–3), then hand the corrected words back to the apprentice's stages. | Problem | Example (heard → wrongly written → fix) | |---|---| | Dialect turned into فصحى | said «إيه» → «ماذا» → «إيه»; «دلوقتي» → «الآن» → «دلوقتي» | | English terms in Arabic letters | «بروميت» → «prompt»; «إيديتنج» → «editing»; «ثمبنيل» → keep as the creator writes it | | Brand names | «كلود» → «Claude»; «كاب كت» → «CapCut» | | Word splits/merges | «ما عرفتش» ↔ «ماعرفتش»: pick one and stay consistent | | Numbers | «تلاتة» vs «3»: words in the transcript, digits in captions (per the Voice Card) | | Hamza / taa marbuta / alif maqsura | normalize to standard spelling (إ/أ, ة, ى) unless it's intentional

This is an excerpt. Refer to the complete source and its current requirements.