Turns LaTeX and Markdown documents into text that a speech model can read aloud. Maths is spoken the way a person would say it. Preamble, comments, code blocks, table cells and the bibliography are not spoken.
Status: the English engine works. tex2speech FILE turns a .tex or .md file into one sentence per line, --json gives segments with their source positions, and - with --format tex|md reads standard input. The benchmark transcripts in bench/ match exactly. Portuguese and French are not read yet. tex2speech --version prints the version.
- A Rust library that other programs link. It takes a whole file, a paragraph or a sentence.
- A program,
tex2speech, that prints plain text or JSON. - No network access and no configuration for the default run.
Rust 1.99 or later.
cargo build --release
cargo test
Apache-2.0. See LICENSE and NOTICE.