Crate litedoc_core 

LiteDoc Core

A deterministic, AI-token-efficient document format parser.

LiteDoc provides an alternative to Markdown with explicit fenced block syntax, making it ideal for AI consumption due to reduced ambiguity and lower token counts.

Quick Start

use litedoc_core::{Parser, Profile};

let input = "# Hello World\n\nThis is a **paragraph**.";
let mut parser = Parser::new(Profile::Litedoc);
let doc = parser.parse(input).unwrap();

println!("Parsed {} blocks", doc.blocks.len());

Error Recovery

The parser supports graceful error recovery:

use litedoc_core::{Parser, Profile};

let input = "::unknown\nsome content\n::";
let mut parser = Parser::new(Profile::Litedoc);
let result = parser.parse_with_recovery(input);

// Document is still parsed, errors are collected
println!("Blocks: {}, Errors: {}", result.document.blocks.len(), result.errors.len());

Profiles

• Profile::Litedoc - Full native syntax with explicit fencing
• Profile::Md - CommonMark + GFM subset
• Profile::MdStrict - CommonMark core only

Re-exports

pub use ast::Block;

pub use ast::Document;

pub use ast::Inline;

pub use ast::Profile;

pub use error::ParseError;

pub use error::ParseErrorKind;

pub use error::ParseErrors;

pub use parser::ParseResult;

pub use parser::Parser;

Modules

ast

Abstract Syntax Tree types for LiteDoc documents.

error

inline

Zero-allocation inline parser for LiteDoc

lexer

Line-based lexer with SIMD-accelerated scanning.

parser

Zero-allocation block parser for LiteDoc

span

Source location tracking for AST nodes.