Skip to main content
The Worlds documentation follows a strict, highly-opinionated set of editorial guidelines. These guidelines optimize for developer momentum, absolute technical authority, and zero marketing fluff.

File conventions

  • Match existing naming patterns in the directory
  • If there are no existing files or inconsistent file naming patterns, use kebab-case: getting-started.mdx, api-reference.mdx

Architecture & structural integrity

  • Quality over Quantity: Prioritize concision and ruthless word economy. Every word must earn its keep. Strip stylistic adjectives (e.g., “these logical, structured facts” -> “facts”) and reduce verbose transitional phrases (e.g., “transition cleanly into using” -> “switch to”). If a sentence delivers the same technical weight in fewer words, it must be rewritten. Eliminate filler sections and ensure every page earns its place.
    • Bad: “By combining these logical, structured facts with a search index, your agent can use semantic search to instantly find the right item, and then transition cleanly into using the RDF graph to reason about everything connected to it.”
    • Good: “By combining facts with a search index, agents can use semantic search to instantly find the right item, and then switch to reasoning over the RDF graph.”
  • Actionable Documentation: Publish pages only when they contain actionable technical value. Document what exists, or omit the page entirely.

The “Code-Forward” pattern

Developers come to the docs to copy code. Give it to them immediately.
  • Brief Opener: Start the page with a maximum of two sentences defining what the entity is, using active voice.
  • Immediate <CodeGroup>: The very next element must be a multi-language code snippet, including TypeScript, Python, and cURL, showing the entity in action.
  • API at a Glance: Follow the code block with a Markdown table summarizing the core methods.
  • Deep Dive Below the Fold: Place architectural explanations, concepts, and diagrams below the usable code and API tables.

Writing standards

Voice and structure

  • Second-person active voice: Use “you” (“You can write…” rather than “Developers can write…”). Active voice, direct language.
  • Positive Phrasing: State what to do rather than what to avoid. Affirmative framing aids comprehension.
  • Prescriptive Guidance: Tell the user exactly what to do. Use definitive verbs like must, requires, or use.
  • Timeless Documentation: Write as if the reader is viewing the docs years from now.
  • Global & Inclusive Language: Write for a global audience. Omit idioms, slang, and exclusionary terms.
  • Sentence-case headings: Use “Getting started” and not “Getting Started”. Keep headings concise in as few words as possible. Avoid colons.
  • Sentence-case code titles: Use “Expandable example” and not “Expandable Example”.
  • Lead with context: Explain what something is before how to use it.
  • Prerequisites first: Place prerequisites at the start of procedural content.
  • Keep markers in body: Keep contextual markers (like dates) in the body text, not the heading.
  • Contextual examples: Avoid repeatedly calling back to specific examples or analogies later in a document if they lose context.

What to avoid

  • Zero Marketing Adjectives: Avoid words like “powerful”, “seamless”, “robust”, or “cutting-edge” in reference pages.
  • Filler phrases: Omit transitional cliches like “at its core”, “it’s important to note”, or “in order to”. Use direct statements.
  • Punctuation cliches: Do not use em dashes (—) as a crutch to connect loose thoughts. Terminate the thought and write a new, concise sentence.
  • Excessive conjunctions: Omit “moreover”, “furthermore”, or “additionally”.
  • Editorializing words: Omit “obviously”, “simply”, “just”, or “easily”. Acknowledging complexity respects the user.
  • Generic fluff: Omit generic introductions that don’t add value, and concluding summaries that restate what was just said.
  • Imaginative phrasing: Avoid anthropomorphic language to describe AI capabilities (e.g., avoid “intuition”, “brain”, or “thinking”; use “understanding”, “model”, or “processing”).
  • Idioms and metaphors: Avoid phrases that obscure technical precision (e.g., avoid “black box”, “heavy lifting”, or “plumbing”; use literal descriptions).
  • Single-model references: Discuss general AI capabilities with future-proof lists (e.g., “Gemini, Claude, or ChatGPT” instead of just “GPT-4”).

Formatting

  • Parentheses: Use sparingly and only for direct definitions, such as acronyms. Otherwise, revise parenthetical thoughts into fluid, formal sentences.
  • Language tags: All code blocks must have explicit language tags.
  • Alt text: All images and media must have descriptive alt text.
  • Link text: Use descriptive text for links rather than generic phrases like “click here”.
  • Semantic styling: Use bold and italics only when they serve the reader’s understanding—never use text styling just for decoration.
  • No decorative emojis: No decorative formatting or emoji (exception: decorative emojis are permitted in footer sections).
  • Ordered Lists: Use ordered or numbered lists specifically for deliberate sequences, step-by-step procedures, or strict hierarchies. Default to unordered bullet points.

Code examples

  • Source-Truth Alignment: Ensure every method signature and property in code blocks perfectly mirrors the TypeScript definitions.
  • Keep it simple: Keep examples practical and simple.
  • Realistic values: Use realistic values rather than generic names like “foo” or “bar”.
  • Singular focus: One clear example is better than multiple variations.
  • Tested code: Test that code works before including it.

Sources and citations