Lacko Audio Generator — a multilingual TTS pipeline that replaces a recording studio. a multilingual TTS pipeline that replaces a recording studio.

A Python batch pipeline that reads a CSV of voice-line definitions and produces ~80+ MP3s in three languages (Slovak, English, Hungarian) using the ElevenLabs API — feeds the Lacko app.

Output organised by chapter (CH01–CH10) and content type (story / learn / minigame). One CSV edit, one rerun, fresh audio across all languages — removes the recording-studio bottleneck for localising the museum app.

Hard parts. Rate-limit and cost control on the ElevenLabs API. Voice selection and prosody tuning per language. CSV encoding for Slovak / Hungarian diacritics on a Windows-default-cp1252 console.