pdf-oxide-wasm 0.3.51 → 0.3.52

package/bundler/pdf_oxide.d.ts

@@ -644,7 +644,9 @@ export class WasmHeader {
 }
 
 /**
- * OCR configuration for WebAssembly.
+ * OCR configuration for WebAssembly. (Currently a marker — the engine
+ * uses tuned defaults; knobs are exposed as the WASM OCR surface
+ * matures, #524.)
  */
 export class WasmOcrConfig {
     free(): void;
@@ -656,15 +658,27 @@ export class WasmOcrConfig {
 }
 
 /**
- * OCR engine for WebAssembly.
+ * OCR engine for WebAssembly (#524).
+ *
+ * OCR runs entirely in-WASM via the pure-Rust `tract` backend — no
+ * native ONNX Runtime, no JS bridge. Model **delivery is host-side**:
+ * the browser/Deno/edge host fetches the detector + recognizer ONNX
+ * files and the char dictionary (see `modelManifest()` for the URLs)
+ * — typically `fetch()` + the Cache API / IndexedDB for the
+ * tens-of-MB models — then hands the bytes to the constructor. This
+ * only works in the `wasm-ocr` build of `pdf-oxide`; the default
+ * `pdf-oxide-wasm` has no OCR (the constructor returns an error
+ * explaining this).
  */
 export class WasmOcrEngine {
     free(): void;
     [Symbol.dispose](): void;
     /**
-     * Create a new OCR engine.
+     * Not available in this build. OCR needs the `wasm-ocr` build of
+     * `pdf-oxide` (the pure-Rust tract backend); the default
+     * `pdf-oxide-wasm` ships without it.
      */
-    constructor(_det_model_path: string, _rec_model_path: string, _dict_path: string, _config?: WasmOcrConfig | null);
+    constructor(_det_model: Uint8Array, _rec_model: Uint8Array, _dict: string, _config?: WasmOcrConfig | null);
 }
 
 /**
@@ -1017,12 +1031,10 @@ export class WasmPdfDocument {
      */
     extractTextLines(page_index: number, region?: Float32Array | null): any;
     /**
-     * Extract text using OCR (optical character recognition).
-     *
-     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
-     * ONNX Runtime support for the web backend in the current implementation.
+     * Extract text using OCR. Not available in this build — OCR needs
+     * the `wasm-ocr` build of `pdf-oxide`.
      */
-    extractTextOcr(_page_index: number, _engine?: WasmOcrEngine | null): string;
+    extractTextOcr(_page_index: number, _engine: WasmOcrEngine): string;
     /**
      * Extract word-level data from a page.
      *
@@ -1445,6 +1457,10 @@ export class WasmPdfPageRegion {
     extractTextLines(): any;
     /**
      * Extract text using OCR from this region.
+     *
+     * Region-scoped OCR is not wired yet; use the page-level
+     * `WasmPdfDocument.extractTextOcr(pageIndex, engine)` for now
+     * (#524 follow-up).
      */
     extractTextOcr(_engine?: WasmOcrEngine | null): string;
     /**

package/bundler/pdf_oxide_bg.js

@@ -2544,7 +2544,9 @@ export class WasmHeader {
 if (Symbol.dispose) WasmHeader.prototype[Symbol.dispose] = WasmHeader.prototype.free;
 
 /**
- * OCR configuration for WebAssembly.
+ * OCR configuration for WebAssembly. (Currently a marker — the engine
+ * uses tuned defaults; knobs are exposed as the WASM OCR surface
+ * matures, #524.)
  */
 export class WasmOcrConfig {
     __destroy_into_raw() {
@@ -2570,7 +2572,17 @@ export class WasmOcrConfig {
 if (Symbol.dispose) WasmOcrConfig.prototype[Symbol.dispose] = WasmOcrConfig.prototype.free;
 
 /**
- * OCR engine for WebAssembly.
+ * OCR engine for WebAssembly (#524).
+ *
+ * OCR runs entirely in-WASM via the pure-Rust `tract` backend — no
+ * native ONNX Runtime, no JS bridge. Model **delivery is host-side**:
+ * the browser/Deno/edge host fetches the detector + recognizer ONNX
+ * files and the char dictionary (see `modelManifest()` for the URLs)
+ * — typically `fetch()` + the Cache API / IndexedDB for the
+ * tens-of-MB models — then hands the bytes to the constructor. This
+ * only works in the `wasm-ocr` build of `pdf-oxide`; the default
+ * `pdf-oxide-wasm` has no OCR (the constructor returns an error
+ * explaining this).
  */
 export class WasmOcrEngine {
     __destroy_into_raw() {
@@ -2584,20 +2596,22 @@ export class WasmOcrEngine {
         wasm.__wbg_wasmocrengine_free(ptr, 0);
     }
     /**
-     * Create a new OCR engine.
-     * @param {string} _det_model_path
-     * @param {string} _rec_model_path
-     * @param {string} _dict_path
+     * Not available in this build. OCR needs the `wasm-ocr` build of
+     * `pdf-oxide` (the pure-Rust tract backend); the default
+     * `pdf-oxide-wasm` ships without it.
+     * @param {Uint8Array} _det_model
+     * @param {Uint8Array} _rec_model
+     * @param {string} _dict
      * @param {WasmOcrConfig | null} [_config]
      */
-    constructor(_det_model_path, _rec_model_path, _dict_path, _config) {
+    constructor(_det_model, _rec_model, _dict, _config) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            const ptr0 = passStringToWasm0(_det_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr0 = passArray8ToWasm0(_det_model, wasm.__wbindgen_export);
             const len0 = WASM_VECTOR_LEN;
-            const ptr1 = passStringToWasm0(_rec_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr1 = passArray8ToWasm0(_rec_model, wasm.__wbindgen_export);
             const len1 = WASM_VECTOR_LEN;
-            const ptr2 = passStringToWasm0(_dict_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr2 = passStringToWasm0(_dict, wasm.__wbindgen_export, wasm.__wbindgen_export2);
             const len2 = WASM_VECTOR_LEN;
             let ptr3 = 0;
             if (!isLikeNone(_config)) {
@@ -3902,41 +3916,35 @@ export class WasmPdfDocument {
         }
     }
     /**
-     * Extract text using OCR (optical character recognition).
-     *
-     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
-     * ONNX Runtime support for the web backend in the current implementation.
+     * Extract text using OCR. Not available in this build — OCR needs
+     * the `wasm-ocr` build of `pdf-oxide`.
      * @param {number} _page_index
-     * @param {WasmOcrEngine | null} [_engine]
+     * @param {WasmOcrEngine} _engine
      * @returns {string}
      */
     extractTextOcr(_page_index, _engine) {
-        let deferred3_0;
-        let deferred3_1;
+        let deferred2_0;
+        let deferred2_1;
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            let ptr0 = 0;
-            if (!isLikeNone(_engine)) {
-                _assertClass(_engine, WasmOcrEngine);
-                ptr0 = _engine.__destroy_into_raw();
-            }
-            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, ptr0);
+            _assertClass(_engine, WasmOcrEngine);
+            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, _engine.__wbg_ptr);
             var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
             var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
             var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
             var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
-            var ptr2 = r0;
-            var len2 = r1;
+            var ptr1 = r0;
+            var len1 = r1;
             if (r3) {
-                ptr2 = 0; len2 = 0;
+                ptr1 = 0; len1 = 0;
                 throw takeObject(r2);
             }
-            deferred3_0 = ptr2;
-            deferred3_1 = len2;
-            return getStringFromWasm0(ptr2, len2);
+            deferred2_0 = ptr1;
+            deferred2_1 = len1;
+            return getStringFromWasm0(ptr1, len1);
         } finally {
             wasm.__wbindgen_add_to_stack_pointer(16);
-            wasm.__wbindgen_export4(deferred3_0, deferred3_1, 1);
+            wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
         }
     }
     /**
@@ -5695,6 +5703,10 @@ export class WasmPdfPageRegion {
     }
     /**
      * Extract text using OCR from this region.
+     *
+     * Region-scoped OCR is not wired yet; use the page-level
+     * `WasmPdfDocument.extractTextOcr(pageIndex, engine)` for now
+     * (#524 follow-up).
      * @param {WasmOcrEngine | null} [_engine]
      * @returns {string}
      */

package/nodejs/pdf_oxide.d.ts

@@ -644,7 +644,9 @@ export class WasmHeader {
 }
 
 /**
- * OCR configuration for WebAssembly.
+ * OCR configuration for WebAssembly. (Currently a marker — the engine
+ * uses tuned defaults; knobs are exposed as the WASM OCR surface
+ * matures, #524.)
  */
 export class WasmOcrConfig {
     free(): void;
@@ -656,15 +658,27 @@ export class WasmOcrConfig {
 }
 
 /**
- * OCR engine for WebAssembly.
+ * OCR engine for WebAssembly (#524).
+ *
+ * OCR runs entirely in-WASM via the pure-Rust `tract` backend — no
+ * native ONNX Runtime, no JS bridge. Model **delivery is host-side**:
+ * the browser/Deno/edge host fetches the detector + recognizer ONNX
+ * files and the char dictionary (see `modelManifest()` for the URLs)
+ * — typically `fetch()` + the Cache API / IndexedDB for the
+ * tens-of-MB models — then hands the bytes to the constructor. This
+ * only works in the `wasm-ocr` build of `pdf-oxide`; the default
+ * `pdf-oxide-wasm` has no OCR (the constructor returns an error
+ * explaining this).
  */
 export class WasmOcrEngine {
     free(): void;
     [Symbol.dispose](): void;
     /**
-     * Create a new OCR engine.
+     * Not available in this build. OCR needs the `wasm-ocr` build of
+     * `pdf-oxide` (the pure-Rust tract backend); the default
+     * `pdf-oxide-wasm` ships without it.
      */
-    constructor(_det_model_path: string, _rec_model_path: string, _dict_path: string, _config?: WasmOcrConfig | null);
+    constructor(_det_model: Uint8Array, _rec_model: Uint8Array, _dict: string, _config?: WasmOcrConfig | null);
 }
 
 /**
@@ -1017,12 +1031,10 @@ export class WasmPdfDocument {
      */
     extractTextLines(page_index: number, region?: Float32Array | null): any;
     /**
-     * Extract text using OCR (optical character recognition).
-     *
-     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
-     * ONNX Runtime support for the web backend in the current implementation.
+     * Extract text using OCR. Not available in this build — OCR needs
+     * the `wasm-ocr` build of `pdf-oxide`.
      */
-    extractTextOcr(_page_index: number, _engine?: WasmOcrEngine | null): string;
+    extractTextOcr(_page_index: number, _engine: WasmOcrEngine): string;
     /**
      * Extract word-level data from a page.
      *
@@ -1445,6 +1457,10 @@ export class WasmPdfPageRegion {
     extractTextLines(): any;
     /**
      * Extract text using OCR from this region.
+     *
+     * Region-scoped OCR is not wired yet; use the page-level
+     * `WasmPdfDocument.extractTextOcr(pageIndex, engine)` for now
+     * (#524 follow-up).
      */
     extractTextOcr(_engine?: WasmOcrEngine | null): string;
     /**

package/nodejs/pdf_oxide.js

@@ -2559,7 +2559,9 @@ if (Symbol.dispose) WasmHeader.prototype[Symbol.dispose] = WasmHeader.prototype.
 exports.WasmHeader = WasmHeader;
 
 /**
- * OCR configuration for WebAssembly.
+ * OCR configuration for WebAssembly. (Currently a marker — the engine
+ * uses tuned defaults; knobs are exposed as the WASM OCR surface
+ * matures, #524.)
  */
 class WasmOcrConfig {
     __destroy_into_raw() {
@@ -2586,7 +2588,17 @@ if (Symbol.dispose) WasmOcrConfig.prototype[Symbol.dispose] = WasmOcrConfig.prot
 exports.WasmOcrConfig = WasmOcrConfig;
 
 /**
- * OCR engine for WebAssembly.
+ * OCR engine for WebAssembly (#524).
+ *
+ * OCR runs entirely in-WASM via the pure-Rust `tract` backend — no
+ * native ONNX Runtime, no JS bridge. Model **delivery is host-side**:
+ * the browser/Deno/edge host fetches the detector + recognizer ONNX
+ * files and the char dictionary (see `modelManifest()` for the URLs)
+ * — typically `fetch()` + the Cache API / IndexedDB for the
+ * tens-of-MB models — then hands the bytes to the constructor. This
+ * only works in the `wasm-ocr` build of `pdf-oxide`; the default
+ * `pdf-oxide-wasm` has no OCR (the constructor returns an error
+ * explaining this).
  */
 class WasmOcrEngine {
     __destroy_into_raw() {
@@ -2600,20 +2612,22 @@ class WasmOcrEngine {
         wasm.__wbg_wasmocrengine_free(ptr, 0);
     }
     /**
-     * Create a new OCR engine.
-     * @param {string} _det_model_path
-     * @param {string} _rec_model_path
-     * @param {string} _dict_path
+     * Not available in this build. OCR needs the `wasm-ocr` build of
+     * `pdf-oxide` (the pure-Rust tract backend); the default
+     * `pdf-oxide-wasm` ships without it.
+     * @param {Uint8Array} _det_model
+     * @param {Uint8Array} _rec_model
+     * @param {string} _dict
      * @param {WasmOcrConfig | null} [_config]
      */
-    constructor(_det_model_path, _rec_model_path, _dict_path, _config) {
+    constructor(_det_model, _rec_model, _dict, _config) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            const ptr0 = passStringToWasm0(_det_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr0 = passArray8ToWasm0(_det_model, wasm.__wbindgen_export);
             const len0 = WASM_VECTOR_LEN;
-            const ptr1 = passStringToWasm0(_rec_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr1 = passArray8ToWasm0(_rec_model, wasm.__wbindgen_export);
             const len1 = WASM_VECTOR_LEN;
-            const ptr2 = passStringToWasm0(_dict_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr2 = passStringToWasm0(_dict, wasm.__wbindgen_export, wasm.__wbindgen_export2);
             const len2 = WASM_VECTOR_LEN;
             let ptr3 = 0;
             if (!isLikeNone(_config)) {
@@ -3921,41 +3935,35 @@ class WasmPdfDocument {
         }
     }
     /**
-     * Extract text using OCR (optical character recognition).
-     *
-     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
-     * ONNX Runtime support for the web backend in the current implementation.
+     * Extract text using OCR. Not available in this build — OCR needs
+     * the `wasm-ocr` build of `pdf-oxide`.
      * @param {number} _page_index
-     * @param {WasmOcrEngine | null} [_engine]
+     * @param {WasmOcrEngine} _engine
      * @returns {string}
      */
     extractTextOcr(_page_index, _engine) {
-        let deferred3_0;
-        let deferred3_1;
+        let deferred2_0;
+        let deferred2_1;
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            let ptr0 = 0;
-            if (!isLikeNone(_engine)) {
-                _assertClass(_engine, WasmOcrEngine);
-                ptr0 = _engine.__destroy_into_raw();
-            }
-            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, ptr0);
+            _assertClass(_engine, WasmOcrEngine);
+            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, _engine.__wbg_ptr);
             var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
             var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
             var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
             var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
-            var ptr2 = r0;
-            var len2 = r1;
+            var ptr1 = r0;
+            var len1 = r1;
             if (r3) {
-                ptr2 = 0; len2 = 0;
+                ptr1 = 0; len1 = 0;
                 throw takeObject(r2);
             }
-            deferred3_0 = ptr2;
-            deferred3_1 = len2;
-            return getStringFromWasm0(ptr2, len2);
+            deferred2_0 = ptr1;
+            deferred2_1 = len1;
+            return getStringFromWasm0(ptr1, len1);
         } finally {
             wasm.__wbindgen_add_to_stack_pointer(16);
-            wasm.__wbindgen_export4(deferred3_0, deferred3_1, 1);
+            wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
         }
     }
     /**
@@ -5715,6 +5723,10 @@ class WasmPdfPageRegion {
     }
     /**
      * Extract text using OCR from this region.
+     *
+     * Region-scoped OCR is not wired yet; use the page-level
+     * `WasmPdfDocument.extractTextOcr(pageIndex, engine)` for now
+     * (#524 follow-up).
      * @param {WasmOcrEngine | null} [_engine]
      * @returns {string}
      */

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pdf-oxide-wasm",
-  "version": "0.3.51",
+  "version": "0.3.52",
   "description": "Fast, zero-dependency PDF toolkit for Node.js, browsers, and edge runtimes — text extraction, markdown/HTML conversion, search, form filling, creation, and editing. Rust core compiled to WebAssembly.",
   "license": "MIT OR Apache-2.0",
   "repository": {

package/web/pdf_oxide.d.ts

@@ -644,7 +644,9 @@ export class WasmHeader {
 }
 
 /**
- * OCR configuration for WebAssembly.
+ * OCR configuration for WebAssembly. (Currently a marker — the engine
+ * uses tuned defaults; knobs are exposed as the WASM OCR surface
+ * matures, #524.)
  */
 export class WasmOcrConfig {
     free(): void;
@@ -656,15 +658,27 @@ export class WasmOcrConfig {
 }
 
 /**
- * OCR engine for WebAssembly.
+ * OCR engine for WebAssembly (#524).
+ *
+ * OCR runs entirely in-WASM via the pure-Rust `tract` backend — no
+ * native ONNX Runtime, no JS bridge. Model **delivery is host-side**:
+ * the browser/Deno/edge host fetches the detector + recognizer ONNX
+ * files and the char dictionary (see `modelManifest()` for the URLs)
+ * — typically `fetch()` + the Cache API / IndexedDB for the
+ * tens-of-MB models — then hands the bytes to the constructor. This
+ * only works in the `wasm-ocr` build of `pdf-oxide`; the default
+ * `pdf-oxide-wasm` has no OCR (the constructor returns an error
+ * explaining this).
  */
 export class WasmOcrEngine {
     free(): void;
     [Symbol.dispose](): void;
     /**
-     * Create a new OCR engine.
+     * Not available in this build. OCR needs the `wasm-ocr` build of
+     * `pdf-oxide` (the pure-Rust tract backend); the default
+     * `pdf-oxide-wasm` ships without it.
      */
-    constructor(_det_model_path: string, _rec_model_path: string, _dict_path: string, _config?: WasmOcrConfig | null);
+    constructor(_det_model: Uint8Array, _rec_model: Uint8Array, _dict: string, _config?: WasmOcrConfig | null);
 }
 
 /**
@@ -1017,12 +1031,10 @@ export class WasmPdfDocument {
      */
     extractTextLines(page_index: number, region?: Float32Array | null): any;
     /**
-     * Extract text using OCR (optical character recognition).
-     *
-     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
-     * ONNX Runtime support for the web backend in the current implementation.
+     * Extract text using OCR. Not available in this build — OCR needs
+     * the `wasm-ocr` build of `pdf-oxide`.
      */
-    extractTextOcr(_page_index: number, _engine?: WasmOcrEngine | null): string;
+    extractTextOcr(_page_index: number, _engine: WasmOcrEngine): string;
     /**
      * Extract word-level data from a page.
      *
@@ -1445,6 +1457,10 @@ export class WasmPdfPageRegion {
     extractTextLines(): any;
     /**
      * Extract text using OCR from this region.
+     *
+     * Region-scoped OCR is not wired yet; use the page-level
+     * `WasmPdfDocument.extractTextOcr(pageIndex, engine)` for now
+     * (#524 follow-up).
      */
     extractTextOcr(_engine?: WasmOcrEngine | null): string;
     /**

package/web/pdf_oxide.js

@@ -2546,7 +2546,9 @@ export class WasmHeader {
 if (Symbol.dispose) WasmHeader.prototype[Symbol.dispose] = WasmHeader.prototype.free;
 
 /**
- * OCR configuration for WebAssembly.
+ * OCR configuration for WebAssembly. (Currently a marker — the engine
+ * uses tuned defaults; knobs are exposed as the WASM OCR surface
+ * matures, #524.)
  */
 export class WasmOcrConfig {
     __destroy_into_raw() {
@@ -2572,7 +2574,17 @@ export class WasmOcrConfig {
 if (Symbol.dispose) WasmOcrConfig.prototype[Symbol.dispose] = WasmOcrConfig.prototype.free;
 
 /**
- * OCR engine for WebAssembly.
+ * OCR engine for WebAssembly (#524).
+ *
+ * OCR runs entirely in-WASM via the pure-Rust `tract` backend — no
+ * native ONNX Runtime, no JS bridge. Model **delivery is host-side**:
+ * the browser/Deno/edge host fetches the detector + recognizer ONNX
+ * files and the char dictionary (see `modelManifest()` for the URLs)
+ * — typically `fetch()` + the Cache API / IndexedDB for the
+ * tens-of-MB models — then hands the bytes to the constructor. This
+ * only works in the `wasm-ocr` build of `pdf-oxide`; the default
+ * `pdf-oxide-wasm` has no OCR (the constructor returns an error
+ * explaining this).
  */
 export class WasmOcrEngine {
     __destroy_into_raw() {
@@ -2586,20 +2598,22 @@ export class WasmOcrEngine {
         wasm.__wbg_wasmocrengine_free(ptr, 0);
     }
     /**
-     * Create a new OCR engine.
-     * @param {string} _det_model_path
-     * @param {string} _rec_model_path
-     * @param {string} _dict_path
+     * Not available in this build. OCR needs the `wasm-ocr` build of
+     * `pdf-oxide` (the pure-Rust tract backend); the default
+     * `pdf-oxide-wasm` ships without it.
+     * @param {Uint8Array} _det_model
+     * @param {Uint8Array} _rec_model
+     * @param {string} _dict
      * @param {WasmOcrConfig | null} [_config]
      */
-    constructor(_det_model_path, _rec_model_path, _dict_path, _config) {
+    constructor(_det_model, _rec_model, _dict, _config) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            const ptr0 = passStringToWasm0(_det_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr0 = passArray8ToWasm0(_det_model, wasm.__wbindgen_export);
             const len0 = WASM_VECTOR_LEN;
-            const ptr1 = passStringToWasm0(_rec_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr1 = passArray8ToWasm0(_rec_model, wasm.__wbindgen_export);
             const len1 = WASM_VECTOR_LEN;
-            const ptr2 = passStringToWasm0(_dict_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const ptr2 = passStringToWasm0(_dict, wasm.__wbindgen_export, wasm.__wbindgen_export2);
             const len2 = WASM_VECTOR_LEN;
             let ptr3 = 0;
             if (!isLikeNone(_config)) {
@@ -3904,41 +3918,35 @@ export class WasmPdfDocument {
         }
     }
     /**
-     * Extract text using OCR (optical character recognition).
-     *
-     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
-     * ONNX Runtime support for the web backend in the current implementation.
+     * Extract text using OCR. Not available in this build — OCR needs
+     * the `wasm-ocr` build of `pdf-oxide`.
      * @param {number} _page_index
-     * @param {WasmOcrEngine | null} [_engine]
+     * @param {WasmOcrEngine} _engine
      * @returns {string}
      */
     extractTextOcr(_page_index, _engine) {
-        let deferred3_0;
-        let deferred3_1;
+        let deferred2_0;
+        let deferred2_1;
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            let ptr0 = 0;
-            if (!isLikeNone(_engine)) {
-                _assertClass(_engine, WasmOcrEngine);
-                ptr0 = _engine.__destroy_into_raw();
-            }
-            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, ptr0);
+            _assertClass(_engine, WasmOcrEngine);
+            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, _engine.__wbg_ptr);
             var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
             var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
             var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
             var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
-            var ptr2 = r0;
-            var len2 = r1;
+            var ptr1 = r0;
+            var len1 = r1;
             if (r3) {
-                ptr2 = 0; len2 = 0;
+                ptr1 = 0; len1 = 0;
                 throw takeObject(r2);
             }
-            deferred3_0 = ptr2;
-            deferred3_1 = len2;
-            return getStringFromWasm0(ptr2, len2);
+            deferred2_0 = ptr1;
+            deferred2_1 = len1;
+            return getStringFromWasm0(ptr1, len1);
         } finally {
             wasm.__wbindgen_add_to_stack_pointer(16);
-            wasm.__wbindgen_export4(deferred3_0, deferred3_1, 1);
+            wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
         }
     }
     /**
@@ -5697,6 +5705,10 @@ export class WasmPdfPageRegion {
     }
     /**
      * Extract text using OCR from this region.
+     *
+     * Region-scoped OCR is not wired yet; use the page-level
+     * `WasmPdfDocument.extractTextOcr(pageIndex, engine)` for now
+     * (#524 follow-up).
      * @param {WasmOcrEngine | null} [_engine]
      * @returns {string}
      */