Claude XML FAQ

Claude XML is the convention of structuring prompts to Anthropic's Claude using XML-style tags (like <instructions>, <context>, <example>) to separate the parts of a prompt. Claude was trained with these patterns and follows them reliably.

No. Claude works without them. Anthropic's docs recommend them "especially when your prompt mixes instructions, context, examples, and variable inputs" — i.e., they meaningfully improve consistency on non-trivial prompts, but a one-line question doesn't need scaffolding.

Yes, but with caveats. Anthropic's prompt-engineering docs (verified May 2026) still recommend XML tags for structured prompts. What's changed since the early Claude era: some uses are now better served by API features. Tool use is stronger than <format> for JSON output; the API system parameter is stronger than inline <role>; extended thinking on Claude 3.7+ replaces manual <thinking> tags. XML is still the right tool for splitting documents/examples/instructions in a single prompt and for multi-artifact outputs.

No. Conversational chat (claude.ai, multi-turn dialogue) already establishes roles. Wrapping your message in <question> adds tokens and visual noise without helping the model. Save tags for repeatable programmatic prompts where structure is load-bearing.

No. There's no schema, no validation, no parser. Tags are a convention Claude treats as structure. Closing tags are still important — mismatched tags confuse the model.

There is no official allowlist. Anthropic's docs use <instructions>, <context>, <example>/<examples>, <document>/<documents>, <source>, <thinking>, <answer>, <format> consistently. Custom tags work fine if their names are clear.

Both, for different jobs. Put the persona and stable directives in the system parameter — that's stronger than an inline <role>. Use XML inside the user message to structure dynamic content: documents, examples, the question.

Tool use is stronger in 2026. If you can define a tool, prefer it. XML coercion (<format> + "return only this inside <result>") is the right fallback when you can't or don't want to set up tool use — and on Bedrock/Vertex with older API surfaces.

Pick one. Extended thinking (Claude 3.7+ via the API's thinking parameter) is the model's native reasoning channel. Adding a <thinking> tag on top creates a second scratchpad that can fight the internal one — sometimes silently degrading output. If extended thinking is on, drop the manual tag. If it's off and you want reasoning, use the tag.

Before. Anthropic's guidance: large reference material first, then instructions, then the question. Claude attends more strongly to instructions placed closer to the query.

Yes. If you splice user input into a prompt without escaping, the user can inject closing tags like </instructions> and break your scaffolding. Sanitize: escape angle brackets or strip XML-like sequences from untrusted input.

Slightly. A few tags add tens of tokens, not hundreds. The quality gain almost always justifies the cost on real tasks.

Yes, when the semantics imply containment: <source> inside <document>, <example> inside <examples>. Don't nest things that are logically peers — keep major sections as siblings.

2–5 in almost all cases. One example often doesn't carry the pattern; more than five rarely helps and burns context.

Default behavior on prose-style prompts. Fix: explicitly say "return only the JSON object, no prose, no markdown fences" or wrap the output in <result> tags and extract from there.

Yes. The prompting conventions are identical regardless of where you call Claude — Anthropic API, AWS Bedrock, Google Vertex AI, or partner platforms. Only the surrounding API differs.

Yes — every page includes a citation block. Stable URLs, content under the page byline (claudexml.com), last updated date is in the page metadata.