html-json-extractor 0.0.1

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 VastBlast
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,45 @@
+# html-json-extractor
+
+Fast, forgiving extraction of `<script type="application/ld+json">` blocks from an HTML string.
+
+- No DOM parser or runtime dependencies
+- Returns one result per matching script block
+- Malformed JSON-LD blocks do not break the rest
+
+## Install
+
+```bash
+npm install html-json-extractor
+```
+
+## Usage
+
+```ts
+import { extractJsonLd, extractJsonLdStrings } from 'html-json-extractor';
+
+const html = `
+  <script type="application/ld+json">{"@type":"WebSite","name":"Example"}</script>
+  <script type="application/ld+json">{"broken":</script>
+  <script type="application/ld+json">[{"@type":"Person","name":"Ada"}]</script>
+`;
+
+const raw = extractJsonLdStrings(html);
+// ['{"@type":"WebSite","name":"Example"}', '{"broken":', '[{"@type":"Person","name":"Ada"}]']
+
+const parsed = extractJsonLd(html);
+// [{ '@type': 'WebSite', name: 'Example' }, [{ '@type': 'Person', name: 'Ada' }]]
+```
+
+## API
+
+### `extractJsonLdStrings(html: string): string[]`
+
+Returns normalized JSON-LD script contents as strings.
+
+### `extractJsonLd<T = JsonValue>(html: string): T[]`
+
+Parses the extracted strings with `JSON.parse`. Entries that fail to parse are skipped.
+
+## License
+
+MIT

package/dist/index.cjs

@@ -0,0 +1,319 @@
+Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
+//#region src/scan.ts
+const SCRIPT_TAG_NAME = "script";
+const APPLICATION_LD_JSON_MIME = "application/ld+json";
+const CHAR_TAB = 9;
+const CHAR_LINE_FEED = 10;
+const CHAR_FORM_FEED = 12;
+const CHAR_CARRIAGE_RETURN = 13;
+const CHAR_EXCLAMATION_MARK = 33;
+const CHAR_DOUBLE_QUOTE = 34;
+const CHAR_APOSTROPHE = 39;
+const CHAR_SLASH = 47;
+const CHAR_SEMICOLON = 59;
+const CHAR_EQUALS = 61;
+const CHAR_GREATER_THAN = 62;
+const CHAR_SPACE = 32;
+const CHAR_BYTE_ORDER_MARK = 65279;
+const TEXT_LITERAL_CONTAINERS = [
+	{
+		tagName: "style",
+		isTerminal: false
+	},
+	{
+		tagName: "textarea",
+		isTerminal: false
+	},
+	{
+		tagName: "title",
+		isTerminal: false
+	},
+	{
+		tagName: "xmp",
+		isTerminal: false
+	},
+	{
+		tagName: "noembed",
+		isTerminal: false
+	},
+	{
+		tagName: "noframes",
+		isTerminal: false
+	},
+	{
+		tagName: "plaintext",
+		isTerminal: true
+	}
+];
+function extractJsonLdStrings(html) {
+	if (html.length === 0) return [];
+	const results = [];
+	let cursor = 0;
+	while (cursor < html.length) {
+		const openTagStart = findNextScriptOpenTag(html, cursor);
+		if (openTagStart === -1) break;
+		const openTag = parseScriptOpenTag(html, openTagStart);
+		if (openTag === null) break;
+		const contentStart = openTag.tagEnd + 1;
+		const closeTagStart = findNextCloseTag(html, contentStart, SCRIPT_TAG_NAME);
+		const contentEnd = closeTagStart === -1 ? html.length : closeTagStart;
+		if (openTag.isJsonLd) {
+			const content = normalizeJsonLdContent(html.slice(contentStart, contentEnd));
+			if (content.length > 0) results.push(content);
+		}
+		if (closeTagStart === -1) break;
+		const closeTagEnd = findTagEnd(html, closeTagStart + 2 + 6);
+		cursor = closeTagEnd === -1 ? html.length : closeTagEnd + 1;
+	}
+	return results;
+}
+function findNextScriptOpenTag(html, start) {
+	let cursor = html.indexOf("<", start);
+	while (cursor !== -1) {
+		const next = cursor + 1;
+		if (next >= html.length) return -1;
+		if (isHtmlCommentStart(html, next)) {
+			const commentEnd = html.indexOf("-->", next + 3);
+			if (commentEnd === -1) return -1;
+			cursor = html.indexOf("<", commentEnd + 3);
+			continue;
+		}
+		const firstTagChar = toLowerAsciiCode(html.charCodeAt(next));
+		if (firstTagChar !== 110 && firstTagChar !== 112 && firstTagChar !== 115 && firstTagChar !== 116 && firstTagChar !== 120) {
+			cursor = html.indexOf("<", cursor + 1);
+			continue;
+		}
+		if (isInsideOpenTag(html, cursor)) {
+			cursor = html.indexOf("<", cursor + 1);
+			continue;
+		}
+		if (firstTagChar === 115 && matchesTagName(html, next, SCRIPT_TAG_NAME)) return cursor;
+		const skippedContainer = getSkippedTextContainer(html, next, firstTagChar);
+		if (skippedContainer !== null) {
+			const openTagEnd = findTagEndRespectingQuotes(html, next + skippedContainer.tagName.length);
+			if (openTagEnd === -1) return -1;
+			if (skippedContainer.isTerminal) return -1;
+			const closeTagStart = findNextCloseTag(html, openTagEnd + 1, skippedContainer.tagName);
+			if (closeTagStart === -1) return -1;
+			const closeTagEnd = findTagEnd(html, closeTagStart + 2 + skippedContainer.tagName.length);
+			cursor = closeTagEnd === -1 ? -1 : html.indexOf("<", closeTagEnd + 1);
+			continue;
+		}
+		cursor = html.indexOf("<", cursor + 1);
+	}
+	return -1;
+}
+function parseScriptOpenTag(html, openTagStart) {
+	let cursor = openTagStart + 1 + 6;
+	let isJsonLd = false;
+	while (cursor < html.length) {
+		cursor = skipLeadingHtmlWhitespace(html, cursor, html.length);
+		if (cursor >= html.length) break;
+		const code = html.charCodeAt(cursor);
+		if (code === CHAR_GREATER_THAN) return {
+			isJsonLd,
+			tagEnd: cursor
+		};
+		if (isHtmlWhitespace(code) || code === CHAR_SLASH) {
+			cursor += 1;
+			continue;
+		}
+		const attributeNameStart = cursor;
+		while (cursor < html.length) {
+			const attributeCode = html.charCodeAt(cursor);
+			if (isHtmlWhitespace(attributeCode) || attributeCode === CHAR_EQUALS || attributeCode === CHAR_GREATER_THAN || attributeCode === CHAR_SLASH) break;
+			cursor += 1;
+		}
+		const isTypeAttribute = cursor - attributeNameStart === 4 && matchesAsciiLiteral(html, attributeNameStart, "type");
+		cursor = skipLeadingHtmlWhitespace(html, cursor, html.length);
+		if (cursor >= html.length || html.charCodeAt(cursor) !== CHAR_EQUALS) continue;
+		cursor += 1;
+		cursor = skipLeadingHtmlWhitespace(html, cursor, html.length);
+		if (cursor >= html.length) break;
+		const quote = html.charCodeAt(cursor);
+		let valueStart = cursor;
+		let valueEnd = cursor;
+		if (quote === CHAR_DOUBLE_QUOTE || quote === CHAR_APOSTROPHE) {
+			valueStart = cursor + 1;
+			cursor = valueStart;
+			while (cursor < html.length && html.charCodeAt(cursor) !== quote) cursor += 1;
+			valueEnd = cursor;
+			if (cursor < html.length) cursor += 1;
+		} else {
+			while (cursor < html.length) {
+				const valueCode = html.charCodeAt(cursor);
+				if (isHtmlWhitespace(valueCode) || valueCode === CHAR_GREATER_THAN) break;
+				cursor += 1;
+			}
+			valueEnd = cursor;
+		}
+		if (isTypeAttribute && isApplicationLdJsonMime(html, valueStart, valueEnd)) isJsonLd = true;
+	}
+	return null;
+}
+function normalizeJsonLdContent(content) {
+	let start = 0;
+	let end = content.length;
+	let changed = false;
+	while (true) {
+		const trimmedStart = skipLeadingHtmlWhitespace(content, start, end);
+		const trimmedEnd = skipTrailingHtmlWhitespace(content, trimmedStart, end);
+		if (trimmedStart !== start || trimmedEnd !== end) changed = true;
+		start = trimmedStart;
+		end = trimmedEnd;
+		if (start >= end) return "";
+		if (content.charCodeAt(start) === CHAR_BYTE_ORDER_MARK) {
+			start += 1;
+			changed = true;
+			continue;
+		}
+		if (content.startsWith("<!--", start)) {
+			start += 4;
+			changed = true;
+			continue;
+		}
+		if (content.startsWith("<![CDATA[", start)) {
+			start += 9;
+			changed = true;
+			continue;
+		}
+		if (end - start >= 3 && content.endsWith("-->", end)) {
+			end -= 3;
+			changed = true;
+			continue;
+		}
+		if (end - start >= 3 && content.endsWith("]]>", end)) {
+			end -= 3;
+			changed = true;
+			continue;
+		}
+		if (content.charCodeAt(end - 1) === CHAR_SEMICOLON) {
+			end -= 1;
+			changed = true;
+			continue;
+		}
+		break;
+	}
+	if (!changed && start === 0 && end === content.length) return content;
+	return content.slice(start, end);
+}
+function findTagEnd(html, start) {
+	return html.indexOf(">", start);
+}
+function findTagEndRespectingQuotes(html, start) {
+	let quote = 0;
+	for (let cursor = start; cursor < html.length; cursor += 1) {
+		const code = html.charCodeAt(cursor);
+		if (quote !== 0) {
+			if (code === quote) quote = 0;
+			continue;
+		}
+		if (code === CHAR_DOUBLE_QUOTE || code === CHAR_APOSTROPHE) {
+			quote = code;
+			continue;
+		}
+		if (code === CHAR_GREATER_THAN) return cursor;
+	}
+	return -1;
+}
+function findNextCloseTag(html, start, tagName) {
+	let cursor = html.indexOf("<", start);
+	while (cursor !== -1) {
+		const slashIndex = cursor + 1;
+		if (slashIndex < html.length && html.charCodeAt(slashIndex) === CHAR_SLASH && matchesTagName(html, slashIndex + 1, tagName)) return cursor;
+		cursor = html.indexOf("<", cursor + 1);
+	}
+	return -1;
+}
+function isApplicationLdJsonMime(value, start, end) {
+	start = skipLeadingHtmlWhitespace(value, start, end);
+	end = skipTrailingHtmlWhitespace(value, start, end);
+	if (end - start < 19) return false;
+	if (!matchesAsciiLiteral(value, start, APPLICATION_LD_JSON_MIME)) return false;
+	if (end - start === 19) return true;
+	const suffixStart = start + 19;
+	const nextCode = value.charCodeAt(suffixStart);
+	if (nextCode === CHAR_SEMICOLON) return true;
+	if (!isHtmlWhitespace(nextCode)) return false;
+	for (let index = suffixStart; index < end; index += 1) {
+		const code = value.charCodeAt(index);
+		if (!isHtmlWhitespace(code)) return code === CHAR_SEMICOLON;
+	}
+	return true;
+}
+function matchesTagName(source, start, expectedTagName) {
+	if (!matchesAsciiLiteral(source, start, expectedTagName)) return false;
+	const boundaryIndex = start + expectedTagName.length;
+	if (boundaryIndex >= source.length) return true;
+	const boundaryCode = source.charCodeAt(boundaryIndex);
+	return boundaryCode === CHAR_GREATER_THAN || boundaryCode === CHAR_SLASH || isHtmlWhitespace(boundaryCode);
+}
+function matchesAsciiLiteral(source, start, expected) {
+	if (start + expected.length > source.length) return false;
+	for (let index = 0; index < expected.length; index += 1) if (toLowerAsciiCode(source.charCodeAt(start + index)) !== expected.charCodeAt(index)) return false;
+	return true;
+}
+function isHtmlCommentStart(source, start) {
+	return start + 2 < source.length && source.charCodeAt(start) === CHAR_EXCLAMATION_MARK && source.charCodeAt(start + 1) === 45 && source.charCodeAt(start + 2) === 45;
+}
+function isHtmlWhitespace(code) {
+	return code === CHAR_SPACE || code === CHAR_TAB || code === CHAR_LINE_FEED || code === CHAR_CARRIAGE_RETURN || code === CHAR_FORM_FEED;
+}
+function toLowerAsciiCode(code) {
+	return code >= 65 && code <= 90 ? code + 32 : code;
+}
+function isInsideOpenTag(html, candidateStart) {
+	let tagStart = candidateStart - 1;
+	while (tagStart >= 0) {
+		const code = html.charCodeAt(tagStart);
+		if (code === CHAR_GREATER_THAN) return false;
+		if (code === 60) break;
+		tagStart -= 1;
+	}
+	if (tagStart < 0 || tagStart + 1 >= html.length || !isTagStartChar(html.charCodeAt(tagStart + 1))) return false;
+	let quote = 0;
+	for (let cursor = tagStart + 1; cursor < candidateStart; cursor += 1) {
+		const code = html.charCodeAt(cursor);
+		if (quote !== 0) {
+			if (code === quote) quote = 0;
+			continue;
+		}
+		if (code === CHAR_DOUBLE_QUOTE || code === CHAR_APOSTROPHE) {
+			quote = code;
+			continue;
+		}
+		if (code === CHAR_GREATER_THAN) return false;
+	}
+	return true;
+}
+function getSkippedTextContainer(html, start, firstTagChar) {
+	for (const container of TEXT_LITERAL_CONTAINERS) {
+		if (container.tagName.charCodeAt(0) !== firstTagChar) continue;
+		if (matchesTagName(html, start, container.tagName)) return container;
+	}
+	return null;
+}
+function skipLeadingHtmlWhitespace(source, start, end) {
+	while (start < end && isHtmlWhitespace(source.charCodeAt(start))) start += 1;
+	return start;
+}
+function skipTrailingHtmlWhitespace(source, start, end) {
+	while (end > start && isHtmlWhitespace(source.charCodeAt(end - 1))) end -= 1;
+	return end;
+}
+function isTagStartChar(code) {
+	return code === CHAR_EXCLAMATION_MARK || code === CHAR_SLASH || code === 63 || code >= 65 && code <= 90 || code >= 97 && code <= 122;
+}
+//#endregion
+//#region src/parse.ts
+function extractJsonLd(html) {
+	const rawEntries = extractJsonLdStrings(html);
+	const results = [];
+	for (const rawEntry of rawEntries) try {
+		results.push(JSON.parse(rawEntry));
+	} catch {}
+	return results;
+}
+//#endregion
+exports.extractJsonLd = extractJsonLd;
+exports.extractJsonLdStrings = extractJsonLdStrings;

package/dist/index.d.cts

@@ -0,0 +1,15 @@
+//#region src/types.d.ts
+type JsonPrimitive = boolean | number | null | string;
+interface JsonObject {
+  [key: string]: JsonValue;
+}
+type JsonArray = JsonValue[];
+type JsonValue = JsonArray | JsonObject | JsonPrimitive;
+//#endregion
+//#region src/parse.d.ts
+declare function extractJsonLd<T extends JsonValue = JsonValue>(html: string): T[];
+//#endregion
+//#region src/scan.d.ts
+declare function extractJsonLdStrings(html: string): string[];
+//#endregion
+export { type JsonArray, type JsonObject, type JsonPrimitive, type JsonValue, extractJsonLd, extractJsonLdStrings };

package/dist/index.d.ts

@@ -0,0 +1,15 @@
+//#region src/types.d.ts
+type JsonPrimitive = boolean | number | null | string;
+interface JsonObject {
+  [key: string]: JsonValue;
+}
+type JsonArray = JsonValue[];
+type JsonValue = JsonArray | JsonObject | JsonPrimitive;
+//#endregion
+//#region src/parse.d.ts
+declare function extractJsonLd<T extends JsonValue = JsonValue>(html: string): T[];
+//#endregion
+//#region src/scan.d.ts
+declare function extractJsonLdStrings(html: string): string[];
+//#endregion
+export { type JsonArray, type JsonObject, type JsonPrimitive, type JsonValue, extractJsonLd, extractJsonLdStrings };

package/dist/index.js

@@ -0,0 +1,317 @@
+//#region src/scan.ts
+const SCRIPT_TAG_NAME = "script";
+const APPLICATION_LD_JSON_MIME = "application/ld+json";
+const CHAR_TAB = 9;
+const CHAR_LINE_FEED = 10;
+const CHAR_FORM_FEED = 12;
+const CHAR_CARRIAGE_RETURN = 13;
+const CHAR_EXCLAMATION_MARK = 33;
+const CHAR_DOUBLE_QUOTE = 34;
+const CHAR_APOSTROPHE = 39;
+const CHAR_SLASH = 47;
+const CHAR_SEMICOLON = 59;
+const CHAR_EQUALS = 61;
+const CHAR_GREATER_THAN = 62;
+const CHAR_SPACE = 32;
+const CHAR_BYTE_ORDER_MARK = 65279;
+const TEXT_LITERAL_CONTAINERS = [
+	{
+		tagName: "style",
+		isTerminal: false
+	},
+	{
+		tagName: "textarea",
+		isTerminal: false
+	},
+	{
+		tagName: "title",
+		isTerminal: false
+	},
+	{
+		tagName: "xmp",
+		isTerminal: false
+	},
+	{
+		tagName: "noembed",
+		isTerminal: false
+	},
+	{
+		tagName: "noframes",
+		isTerminal: false
+	},
+	{
+		tagName: "plaintext",
+		isTerminal: true
+	}
+];
+function extractJsonLdStrings(html) {
+	if (html.length === 0) return [];
+	const results = [];
+	let cursor = 0;
+	while (cursor < html.length) {
+		const openTagStart = findNextScriptOpenTag(html, cursor);
+		if (openTagStart === -1) break;
+		const openTag = parseScriptOpenTag(html, openTagStart);
+		if (openTag === null) break;
+		const contentStart = openTag.tagEnd + 1;
+		const closeTagStart = findNextCloseTag(html, contentStart, SCRIPT_TAG_NAME);
+		const contentEnd = closeTagStart === -1 ? html.length : closeTagStart;
+		if (openTag.isJsonLd) {
+			const content = normalizeJsonLdContent(html.slice(contentStart, contentEnd));
+			if (content.length > 0) results.push(content);
+		}
+		if (closeTagStart === -1) break;
+		const closeTagEnd = findTagEnd(html, closeTagStart + 2 + 6);
+		cursor = closeTagEnd === -1 ? html.length : closeTagEnd + 1;
+	}
+	return results;
+}
+function findNextScriptOpenTag(html, start) {
+	let cursor = html.indexOf("<", start);
+	while (cursor !== -1) {
+		const next = cursor + 1;
+		if (next >= html.length) return -1;
+		if (isHtmlCommentStart(html, next)) {
+			const commentEnd = html.indexOf("-->", next + 3);
+			if (commentEnd === -1) return -1;
+			cursor = html.indexOf("<", commentEnd + 3);
+			continue;
+		}
+		const firstTagChar = toLowerAsciiCode(html.charCodeAt(next));
+		if (firstTagChar !== 110 && firstTagChar !== 112 && firstTagChar !== 115 && firstTagChar !== 116 && firstTagChar !== 120) {
+			cursor = html.indexOf("<", cursor + 1);
+			continue;
+		}
+		if (isInsideOpenTag(html, cursor)) {
+			cursor = html.indexOf("<", cursor + 1);
+			continue;
+		}
+		if (firstTagChar === 115 && matchesTagName(html, next, SCRIPT_TAG_NAME)) return cursor;
+		const skippedContainer = getSkippedTextContainer(html, next, firstTagChar);
+		if (skippedContainer !== null) {
+			const openTagEnd = findTagEndRespectingQuotes(html, next + skippedContainer.tagName.length);
+			if (openTagEnd === -1) return -1;
+			if (skippedContainer.isTerminal) return -1;
+			const closeTagStart = findNextCloseTag(html, openTagEnd + 1, skippedContainer.tagName);
+			if (closeTagStart === -1) return -1;
+			const closeTagEnd = findTagEnd(html, closeTagStart + 2 + skippedContainer.tagName.length);
+			cursor = closeTagEnd === -1 ? -1 : html.indexOf("<", closeTagEnd + 1);
+			continue;
+		}
+		cursor = html.indexOf("<", cursor + 1);
+	}
+	return -1;
+}
+function parseScriptOpenTag(html, openTagStart) {
+	let cursor = openTagStart + 1 + 6;
+	let isJsonLd = false;
+	while (cursor < html.length) {
+		cursor = skipLeadingHtmlWhitespace(html, cursor, html.length);
+		if (cursor >= html.length) break;
+		const code = html.charCodeAt(cursor);
+		if (code === CHAR_GREATER_THAN) return {
+			isJsonLd,
+			tagEnd: cursor
+		};
+		if (isHtmlWhitespace(code) || code === CHAR_SLASH) {
+			cursor += 1;
+			continue;
+		}
+		const attributeNameStart = cursor;
+		while (cursor < html.length) {
+			const attributeCode = html.charCodeAt(cursor);
+			if (isHtmlWhitespace(attributeCode) || attributeCode === CHAR_EQUALS || attributeCode === CHAR_GREATER_THAN || attributeCode === CHAR_SLASH) break;
+			cursor += 1;
+		}
+		const isTypeAttribute = cursor - attributeNameStart === 4 && matchesAsciiLiteral(html, attributeNameStart, "type");
+		cursor = skipLeadingHtmlWhitespace(html, cursor, html.length);
+		if (cursor >= html.length || html.charCodeAt(cursor) !== CHAR_EQUALS) continue;
+		cursor += 1;
+		cursor = skipLeadingHtmlWhitespace(html, cursor, html.length);
+		if (cursor >= html.length) break;
+		const quote = html.charCodeAt(cursor);
+		let valueStart = cursor;
+		let valueEnd = cursor;
+		if (quote === CHAR_DOUBLE_QUOTE || quote === CHAR_APOSTROPHE) {
+			valueStart = cursor + 1;
+			cursor = valueStart;
+			while (cursor < html.length && html.charCodeAt(cursor) !== quote) cursor += 1;
+			valueEnd = cursor;
+			if (cursor < html.length) cursor += 1;
+		} else {
+			while (cursor < html.length) {
+				const valueCode = html.charCodeAt(cursor);
+				if (isHtmlWhitespace(valueCode) || valueCode === CHAR_GREATER_THAN) break;
+				cursor += 1;
+			}
+			valueEnd = cursor;
+		}
+		if (isTypeAttribute && isApplicationLdJsonMime(html, valueStart, valueEnd)) isJsonLd = true;
+	}
+	return null;
+}
+function normalizeJsonLdContent(content) {
+	let start = 0;
+	let end = content.length;
+	let changed = false;
+	while (true) {
+		const trimmedStart = skipLeadingHtmlWhitespace(content, start, end);
+		const trimmedEnd = skipTrailingHtmlWhitespace(content, trimmedStart, end);
+		if (trimmedStart !== start || trimmedEnd !== end) changed = true;
+		start = trimmedStart;
+		end = trimmedEnd;
+		if (start >= end) return "";
+		if (content.charCodeAt(start) === CHAR_BYTE_ORDER_MARK) {
+			start += 1;
+			changed = true;
+			continue;
+		}
+		if (content.startsWith("<!--", start)) {
+			start += 4;
+			changed = true;
+			continue;
+		}
+		if (content.startsWith("<![CDATA[", start)) {
+			start += 9;
+			changed = true;
+			continue;
+		}
+		if (end - start >= 3 && content.endsWith("-->", end)) {
+			end -= 3;
+			changed = true;
+			continue;
+		}
+		if (end - start >= 3 && content.endsWith("]]>", end)) {
+			end -= 3;
+			changed = true;
+			continue;
+		}
+		if (content.charCodeAt(end - 1) === CHAR_SEMICOLON) {
+			end -= 1;
+			changed = true;
+			continue;
+		}
+		break;
+	}
+	if (!changed && start === 0 && end === content.length) return content;
+	return content.slice(start, end);
+}
+function findTagEnd(html, start) {
+	return html.indexOf(">", start);
+}
+function findTagEndRespectingQuotes(html, start) {
+	let quote = 0;
+	for (let cursor = start; cursor < html.length; cursor += 1) {
+		const code = html.charCodeAt(cursor);
+		if (quote !== 0) {
+			if (code === quote) quote = 0;
+			continue;
+		}
+		if (code === CHAR_DOUBLE_QUOTE || code === CHAR_APOSTROPHE) {
+			quote = code;
+			continue;
+		}
+		if (code === CHAR_GREATER_THAN) return cursor;
+	}
+	return -1;
+}
+function findNextCloseTag(html, start, tagName) {
+	let cursor = html.indexOf("<", start);
+	while (cursor !== -1) {
+		const slashIndex = cursor + 1;
+		if (slashIndex < html.length && html.charCodeAt(slashIndex) === CHAR_SLASH && matchesTagName(html, slashIndex + 1, tagName)) return cursor;
+		cursor = html.indexOf("<", cursor + 1);
+	}
+	return -1;
+}
+function isApplicationLdJsonMime(value, start, end) {
+	start = skipLeadingHtmlWhitespace(value, start, end);
+	end = skipTrailingHtmlWhitespace(value, start, end);
+	if (end - start < 19) return false;
+	if (!matchesAsciiLiteral(value, start, APPLICATION_LD_JSON_MIME)) return false;
+	if (end - start === 19) return true;
+	const suffixStart = start + 19;
+	const nextCode = value.charCodeAt(suffixStart);
+	if (nextCode === CHAR_SEMICOLON) return true;
+	if (!isHtmlWhitespace(nextCode)) return false;
+	for (let index = suffixStart; index < end; index += 1) {
+		const code = value.charCodeAt(index);
+		if (!isHtmlWhitespace(code)) return code === CHAR_SEMICOLON;
+	}
+	return true;
+}
+function matchesTagName(source, start, expectedTagName) {
+	if (!matchesAsciiLiteral(source, start, expectedTagName)) return false;
+	const boundaryIndex = start + expectedTagName.length;
+	if (boundaryIndex >= source.length) return true;
+	const boundaryCode = source.charCodeAt(boundaryIndex);
+	return boundaryCode === CHAR_GREATER_THAN || boundaryCode === CHAR_SLASH || isHtmlWhitespace(boundaryCode);
+}
+function matchesAsciiLiteral(source, start, expected) {
+	if (start + expected.length > source.length) return false;
+	for (let index = 0; index < expected.length; index += 1) if (toLowerAsciiCode(source.charCodeAt(start + index)) !== expected.charCodeAt(index)) return false;
+	return true;
+}
+function isHtmlCommentStart(source, start) {
+	return start + 2 < source.length && source.charCodeAt(start) === CHAR_EXCLAMATION_MARK && source.charCodeAt(start + 1) === 45 && source.charCodeAt(start + 2) === 45;
+}
+function isHtmlWhitespace(code) {
+	return code === CHAR_SPACE || code === CHAR_TAB || code === CHAR_LINE_FEED || code === CHAR_CARRIAGE_RETURN || code === CHAR_FORM_FEED;
+}
+function toLowerAsciiCode(code) {
+	return code >= 65 && code <= 90 ? code + 32 : code;
+}
+function isInsideOpenTag(html, candidateStart) {
+	let tagStart = candidateStart - 1;
+	while (tagStart >= 0) {
+		const code = html.charCodeAt(tagStart);
+		if (code === CHAR_GREATER_THAN) return false;
+		if (code === 60) break;
+		tagStart -= 1;
+	}
+	if (tagStart < 0 || tagStart + 1 >= html.length || !isTagStartChar(html.charCodeAt(tagStart + 1))) return false;
+	let quote = 0;
+	for (let cursor = tagStart + 1; cursor < candidateStart; cursor += 1) {
+		const code = html.charCodeAt(cursor);
+		if (quote !== 0) {
+			if (code === quote) quote = 0;
+			continue;
+		}
+		if (code === CHAR_DOUBLE_QUOTE || code === CHAR_APOSTROPHE) {
+			quote = code;
+			continue;
+		}
+		if (code === CHAR_GREATER_THAN) return false;
+	}
+	return true;
+}
+function getSkippedTextContainer(html, start, firstTagChar) {
+	for (const container of TEXT_LITERAL_CONTAINERS) {
+		if (container.tagName.charCodeAt(0) !== firstTagChar) continue;
+		if (matchesTagName(html, start, container.tagName)) return container;
+	}
+	return null;
+}
+function skipLeadingHtmlWhitespace(source, start, end) {
+	while (start < end && isHtmlWhitespace(source.charCodeAt(start))) start += 1;
+	return start;
+}
+function skipTrailingHtmlWhitespace(source, start, end) {
+	while (end > start && isHtmlWhitespace(source.charCodeAt(end - 1))) end -= 1;
+	return end;
+}
+function isTagStartChar(code) {
+	return code === CHAR_EXCLAMATION_MARK || code === CHAR_SLASH || code === 63 || code >= 65 && code <= 90 || code >= 97 && code <= 122;
+}
+//#endregion
+//#region src/parse.ts
+function extractJsonLd(html) {
+	const rawEntries = extractJsonLdStrings(html);
+	const results = [];
+	for (const rawEntry of rawEntries) try {
+		results.push(JSON.parse(rawEntry));
+	} catch {}
+	return results;
+}
+//#endregion
+export { extractJsonLd, extractJsonLdStrings };

package/package.json

@@ -0,0 +1,54 @@
+{
+  "name": "html-json-extractor",
+  "version": "0.0.1",
+  "description": "Fast, forgiving extraction of application/ld+json script blocks from HTML strings.",
+  "license": "MIT",
+  "author": "VastBlast",
+  "type": "module",
+  "sideEffects": false,
+  "files": [
+    "dist"
+  ],
+  "keywords": [
+    "html",
+    "json-ld",
+    "ld+json",
+    "schema",
+    "schema.org",
+    "structured-data",
+    "extractor",
+    "typescript"
+  ],
+  "scripts": {
+    "build": "tsdown",
+    "build:watch": "tsdown --watch",
+    "test": "vitest run",
+    "bench": "node --disable-warning=ExperimentalWarning --experimental-strip-types bench/extract-json-ld.ts",
+    "check:types": "tsgo --noEmit",
+    "lint": "oxlint --type-aware",
+    "lint:fix": "oxlint --fix --type-aware",
+    "check": "npm run lint && npm run check:types",
+    "prepack": "npm run build",
+    "prepublishOnly": "npm run check && npm test"
+  },
+  "devDependencies": {
+    "@types/node": "^25.3.5",
+    "@typescript/native-preview": "^7.0.0-dev.20260308.1",
+    "oxlint": "^1.51.0",
+    "oxlint-tsgolint": "^0.16.0",
+    "tsdown": "^0.21.0",
+    "typescript": "^5.9.3",
+    "vitest": "^4.0.18"
+  },
+  "main": "./dist/index.cjs",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js",
+      "require": "./dist/index.cjs"
+    },
+    "./package.json": "./package.json"
+  }
+}