Crouton Specification

A crouton is a single, atomic fact statement that remains accurate when extracted from context.

To qualify as a crouton, a statement must:

• Contain a complete, verifiable fact
• Be self-contained and require no surrounding context
• Use explicit entity names, not pronouns or references
• State relationships clearly without inference
• Be structured for retrieval, not narrative flow

If a statement fails any of these criteria, it is not a valid crouton.

Croutons use declarative statements. The following structures are allowed:

• Subject verb object. Example: "NRLC.ai provides AI SEO services."
• Subject is object. Example: "Prechunking SEO is an engineering discipline."
• Subject has attribute. Example: "The service includes audit and diagnostic phases."
• Subject performs action at location. Example: "NRLC operates remotely across global markets."
• Subject provides service for audience. Example: "We serve enterprise clients requiring AI visibility."

All structures must use explicit nouns. No pronouns. No implied subjects.

The following patterns are disallowed in croutons:

• Pronouns without clear antecedents. Example: "We provide services." Use: "NRLC.ai provides AI SEO services."
• References to previous sentences. Example: "This approach works." Use: "Prechunking SEO works by shaping content into atomic facts."
• Conditional statements without context. Example: "If needed, contact us." Use: "Contact NRLC.ai for AI SEO consultation."
• Comparative statements without comparison target. Example: "We are better." Use: "NRLC.ai provides more precise AI visibility diagnosis than checklist audits."
• Vague qualifiers. Example: "Many clients succeed." Use: "Enterprise clients using prechunking SEO achieve consistent AI citation rates."
• Narrative connectors. Example: "Therefore, we recommend." Use: "NRLC.ai recommends prechunking SEO for enterprise content systems."

Each disallowed pattern causes facts to mutate or become ambiguous when extracted.

Atomicity means a crouton contains exactly one fact.

A single sentence can contain multiple facts. These must be split into separate croutons.

Example of non-atomic crouton: "NRLC.ai provides AI SEO services and operates remotely."

Correct atomic croutons: "NRLC.ai provides AI SEO services." and "NRLC.ai operates remotely."

Atomicity ensures each fact can be retrieved independently without carrying unnecessary information.

Compound facts are split by logical operators: and, or, but, as well as, in addition to.

Facts that are inseparable are exceptions. Example: "NRLC.ai operates in the United States and United Kingdom." This is atomic because the locations are a single fact about scope.

Croutons must be versioned when facts change.

When a fact is updated, the old crouton is deprecated and a new crouton is published.

Versioning prevents AI systems from retrieving outdated facts.

Deprecated croutons are removed from published content but may exist in training data.

Versioning is tracked through structured data when possible. Otherwise, through content audit cycles.

Time-sensitive facts must include temporal context. Example: "As of 2024, NRLC.ai serves enterprise clients."

Versioning applies to entity facts, service descriptions, and claims about capabilities or outcomes.

Valid crouton: "NRLC.ai provides site audit services for AI and search visibility."

This is atomic, explicit, and self-contained.

Invalid crouton: "We provide these services."

This uses a pronoun and requires context to identify the subject.

Valid crouton: "Site audits explain why visibility breaks down, not just surface-level issues."

This is a complete, standalone fact about what the service does.

Invalid crouton: "This approach is better."

This requires context to identify what approach is being referenced and what it is better than.

• Prechunking SEO Overview
• Core Concepts
• Prechunking Workflow
• Operator Training Course