Train announcement generator: station announcements in a natural voice
Station announcements are a hobby of their own: people recreate them for model railways, train simulator mods, videos, games, school projects or just for fun. All you need is the text and a calm, clear voice that sounds like a public address system.
This guide shows how real announcements are usually built, which voices suit them, how to make the audio sound like a station, and how to voice a whole timetable of announcements offline and free with SuperClonica.
How an announcement is built
Most announcements follow a fixed order: platform, time and operator, destination, calling points, then a safety or service note. For example: “Platform two for the ten fifteen service to Manchester Piccadilly, calling at Stockport and Wilmslow. This train is formed of eight coaches.” Keep each sentence short and repeat the platform and destination at the end of long messages.
Voices that sound like a station
A calm, even female or male narrator with clear diction sounds most like a public address system. Avoid lively or character voices unless you want a parody. Try a few announcer and narrator voices free on their pages with one line of your announcement, up to 100 characters, no account needed.
Write it so it is read correctly
Write times the way they are spoken, “ten fifteen”, and spell out platform numbers. If a station name is read wrongly, write it the way it sounds. Use full stops between parts for clear pauses, and a comma between calling points so the list flows like a real announcement.
Make it sound like a PA system
Voice the announcement with the free SuperClonica engine: install it with python -m pip install superclonica-cpu and run superclonica --text-file announcement.txt --voice announcer.json --lang en --normalize --output platform2.wav. It runs offline with no character limit, and --normalize keeps every announcement at the same loudness. In an audio editor, add your own chime before it, a short reverb for the hall and a slight band-pass filter for the loudspeaker sound.
A whole timetable at once
For a model railway or a simulator, keep the announcements in a table with one row per train and voice them in a loop with the Python API, one file per row. The engine speaks 31 languages, so you can make announcements for German, French, Spanish or Japanese stations too; Chinese is not supported yet.