Tesseract simply cannot power the GCP OCR API because Tesseract sucks super bad and the GCP API is mint.