Quick start

Two minutes, four commands.

Write a document

Create notes.xtxt:

# Project Log

Neural networks are inspired by the brain, loosely.

@task
Title: Ship the reference parser
Status: In Progress
Owner: Subbu
@endtask

@table(chart="bar", x="Month", y="Signups")
Month | Signups
------|--------
Jan   | 20
Feb   | 35
Mar   | 50
@endtable

Check it

xtxt validate notes.xtxt

notes.xtxt: ok (4 blocks)

Errors mean the file cannot be interpreted unambiguously. Warnings mean it parses but something looks wrong. validate exits non-zero only on errors.

Read it

xtxt render notes.xtxt

Renders to the terminal, wrapped and coloured. Add -w 100 to change the wrap column, --no-color to drop the ANSI codes.

Export it

xtxt export notes.xtxt html -o notes.html

The chart becomes inline SVG with no script and no dependency. Add --interactive if you want a reader to hover for values and hide a series.

The part that matters

xtxt extract notes.xtxt

{
  "outline": [{"level": 1, "text": "Project Log", "line": 1}],
  "tasks": [{
    "title": "Ship the reference parser",
    "status": "In Progress",
    "owner": "Subbu",
    "done": false,
    "line": 5
  }],
  "blocks": [],
  "media": [], "links": [], "code": [],
  "words": 12
}

No prose was parsed to produce that. The task was declared, so it is returned — with the line it came from, so anything a program does with it can be traced back to the source.

Next