Reference

The complete technical reference for the current IntentText language and tooling.

Keywords

IntentText supports a compact canonical keyword set plus namespaced extension forms and the Arabic localized keyword names — and no Latin synonym aliases, so every other word you write is reliably your own custom block.

• All Keywords — sortable, filterable keyword table
• Document Identity — canonical title:, summary:, meta:, context:, track: (plus the x-agent: agent:/model: extensions)
• Content — canonical text:, quote:, info: (callout variant set with type: warning|danger|tip|success), code:, image:, link: (plus x-doc:/x-writer: extensions: def:, figure:, contact:, byline:, epigraph:, caption:, footnote:, dedication:)
• Structure — canonical section:, sub:, toc: (plus the x-doc: ref:/deadline:/embed: extensions)
• Tasks — canonical task:, done:, ask:
• Data — canonical headers:, row:, metric: (plus the x-form: input:/output: extensions)
• Agent — canonical step:, gate:, trigger:, decision:, audit:, result:, policy:, context:, ask: (plus the x-agent: orchestration extensions: signal:, memory:, prompt:, tool:, error:, handoff:, wait:, parallel:, retry:, call:, loop:, checkpoint:, import:, export:, progress:, agent:, model:)
• Trust — canonical contract-tier track:, approve:, sign:, freeze:, amendment:, certify:, route:, require: (plus the machine-managed x-trust: history:/revision: blocks)
• Layout — canonical page:, header:, footer:, watermark:, style:, break: (plus the x-layout: font:/divider: and x-doc: signline: extensions)
• Localized (Arabic) Keywords — the 32 Arabic localized keyword names (عنوان→title, مهمة→task, صف→row, …) that round-trip as written; every non-reserved word is a collision-free custom block

Keyword tiers

The canonical keywords are grouped into a small everyday core set plus opt-in profiles, so a plain .it document needs only ~13 keywords and everything else is opt-in. Tiering is contract metadata — the parser still recognizes every keyword, and unknown keywords pass through as custom.

Tier	Count	Keywords
core	13	title summary meta section sub text info quote code image link task done
agent	9	context ask step decision gate trigger result policy audit
contract	8	track approve sign freeze amendment certify route require
data	3	headers row metric
print	7	toc page header footer watermark style break

That is the complete set of 40 canonical keywords. route:/require: declare a document's in-file approval policy and certify: records an authority certification — all three are stable contract-tier keywords. The history: boundary and revision: entries are not tier members: they live below the audit-log boundary as machine-managed x-trust: blocks. The exact, authoritative set is exported from @dotit/core as KEYWORD_TIERS and CANONICAL_KEYWORDS (with CORE_KEYWORDS for the everyday core tier). See the canonical SPEC.

Format version stamp (optional)

A document MAY declare the grammar version it targets with a single comment in the leading header (SPEC §5.2):

// it-format: 1.0
title: Quarterly Report

It is a comment, so it is excluded from every seal hash and round-trips as trivia — adding or changing it never breaks a seal. The parser exposes it as document.version; the feature level inferred from the blocks actually used is document.detectedFeatureLevel (when no stamp is present, version mirrors it). The stamp is advisory self-description for long-term archives — never required, and never a top-level keyword, so it cannot collide with content. Only the header comment block is honored (never a comment in the body or inside a code: block).

Properties

• Pipe Properties — keyword-specific properties
• Style Properties — visual properties available on any block

Systems

• Templates — {{variables}}, each:, merge API
• Query — block queries, document queries, output formats
• Index Files — .it-index shallow architecture
• CLI — every command, every flag