ST-language

ST-language is a plain-text music notation — notes, abstract chords, drums, dynamics, tempo changes, automations (volume, expression, pan, modulation, effect sends, with curved ramps and hairpins), several voices and lyrics — and this is its reference implementation: a pure-Python library (no dependencies) with command-line tools to check songs and export them to MIDI, MusicXML, ABC and MTXT. It also writes grace notes, chord symbols, D.C./D.S. with Coda and Fine, several lyric verses and transposing parts. It is the notation engine of SoundText, usable on its own.

// a blues riff
Tempo: 96 BPM

Bass:
  8: 2c*2 2e*2 2g*2 2a*2 | 2a#*2 2a*2 2g*2 2e*2 |

Piano:
  4: 4C7 | 4F7 |

Install

pip install st-language

Python 3.9 or later, no other packages needed. To work on the library itself, clone this repository and install it in development mode:

git clone https://github.com/openssound/st-language.git
cd st-language
pip install -e .
python -m pytest tests            # the tests and the conformance suite

Command line

st-language check song.st          # syntax errors and warnings (bar checks, lyrics)
st-language midi song.st -o song.mid
st-language musicxml song.st       # -> song.musicxml
st-language abc song.st            # -> song.abc (ABC 2.1)
st-language mtxt song.st           # -> song.mtxt (MTXT 1.0, one event per line)
st-language mtxt take.mtxt         # MTXT -> take.mid
st-language events song.st         # the events as JSON
echo "4: c d e f | 2g 2g" | st2mid - -o melody.mid --instrument Trumpet

stcheck, st2mid, st2musicxml, st2abc and st2mtxt are shortcuts. A file without track headers is a one-track song (choose the instrument with --instrument). Messages are in English, Italian, French or Spanish (--lang, or the ST_LANGUAGE/LANG environment variables).

Python

import st_language as st

events = st.parse("4: c d e f | 2g 2g")           # timed events (beats)
ok, error = st.validate("4: C Am | F G")
warnings = st.check("4: c d e | f g a b |")      # bar 1: 1 quarter note missing

song = st.load_song("song.st")                   # or st.read_song(text)
song = st.Song(name="Demo", tempo_bpm=100)
song.add_track("Bass", "Bass", "4: c*2 g*1 2c*2")
song.add_pattern("Riff", "8: c d e g")
st.to_midi(song, "demo.mid")
st.to_musicxml(song, "demo.musicxml")
st.to_abc(song, "demo.abc")

Specification

The language and the .st song format are defined in the ST-language specification (CC BY 4.0, also in Italian), with a conformance suite for other implementations.

License

GPL-3.0-or-later, © 2026 Sergio Scolaro. The MTXT-derived voice-name table is MIT (see THIRD_PARTY_NOTICES.md).

In italiano: ST-language e' la notazione musicale testuale di SoundText come libreria Python autonoma, senza dipendenze, con i comandi st-language check | midi | musicxml | abc | mtxt | events (e le scorciatoie stcheck, st2mid, st2musicxml, st2abc, st2mtxt). Si installa con pip install st-language. La specifica, anche in italiano, e' in docs/spec/ del repository.