gchq/CyberChef · error · OperationError

Error performing OCR on image. (${err})

Error message

Error performing OCR on image. (${err})

What it means

Catch-all wrapper around the entire Tesseract.js OCR workflow in OpticalCharacterRecognition.run: Web Worker creation (createWorker with the 'eng' language model and WASM core), worker.recognize on a base64 data-URI, and result extraction. Any exception anywhere in that try block is re-thrown as a single OperationError, with the real cause interpolated via ${err}. isWorkerEnvironment and isImage are checked before this block, so this error is specifically about worker/recognize runtime failure.

Source

Thrown at src/core/operations/OpticalCharacterRecognition.mjs:91

                workerPath: `${assetDir}tesseract/worker.min.js`,
                langPath: `${assetDir}tesseract/lang-data`,
                corePath: `${assetDir}tesseract/tesseract-core.wasm.js`,
                logger: progress => {
                    if (isWorkerEnvironment()) {
                        self.sendStatusMessage(`Status: ${progress.status}${progress.status === "recognizing text" ? ` - ${(parseFloat(progress.progress)*100).toFixed(2)}%`: "" }`);
                    }
                }
            });
            self.sendStatusMessage("Finding text...");
            const result = await worker.recognize(image);

            if (showConfidence) {
                return `Confidence: ${result.data.confidence}%\n\n${result.data.text}`;
            } else {
                return result.data.text;
            }
        } catch (err) {
            throw new OperationError(`Error performing OCR on image. (${err})`);
        }
    }
}

export default OpticalCharacterRecognition;

View on GitHub (pinned to 4290ea7539)

Solutions

1. Read the interpolated ${err}: a 'Failed to fetch' / 404 points at missing tesseract assets - verify ${self.docURL}/assets/tesseract/{worker.min.js,lang-data,tesseract-core.wasm.js} are served and reachable.
2. If err mentions WASM, confirm WebAssembly is enabled in the browser and the .wasm.js core loads.
3. Re-export or re-drop the image and confirm it opens in an image viewer (isImage only checks magic bytes, not full integrity).
4. Try a different OCR Engine Mode argument (Tesseract only / LSTM only / Combined) to isolate engine-specific failures.

import { isImage } from '../lib/FileType.mjs';
if (!isImage(input)) {
    throw new Error('Input is not a supported image (jpg/png/pbm/bmp)');
}

try {
    const text = await op.run(inputBuf, [true, 'LSTM only']);
} catch (e) {
    if (e instanceof OperationError && /Error performing OCR/.test(e.message)) {
        // inner cause is appended after 'image. ('
        console.error('OCR failed:', e.message);
    } else throw e;
}

When it happens

Trigger: tesseract.js fails to fetch worker.min.js, lang-data/eng.traineddata, or tesseract-core.wasm.js from the assetDir (${self.docURL}/assets/tesseract/); the input ArrayBuffer passes isImage (magic bytes ok) but the pixel data is corrupt so worker.recognize rejects; WebAssembly is unavailable or blocked by CSP/policy; the worker times out on a very large image.

Common situations: Running CyberChef offline/airgapped where the asset directory is unreachable; self.docURL misconfigured so asset URLs 404; a browser with WASM disabled; feeding a valid-looking but truncated image; tesseract.js version drift where the bundled worker/core/lang assets no longer match the installed package.

Related errors

• Unsupported file type (supported: jpg,png,pbm,bmp) or no fil
• Error opening image. (${err})
• Error normalising image. (${err})
• Could not read a QR code from the image.
• Invalid file type.

AI-assisted analysis of gchq/CyberChef@4290ea7539 (2026-08-13). Data as JSON: /api/errors/7f579e15dd6209e9. Report an issue: GitHub.