Can I upload multiple images at once for bulk OCR?
The workbench processes one image per job. For true batch automation, use the REST API: submit jobs programmatically, one image per request, and collect structured JSON results as they complete. Saved preferences standardize the model across every job in the batch.
Which models are most efficient for bulk image to text?
For high-volume batches, low-cost models keep per-image credit use predictable: PaddleOCR, Gemini 3 Flash, and GPT-5.4 mini each cost 1 credit per document. Compare them on a representative sample first — the right model depends on your image type, not a general ranking.
Is there an API for bulk image to text automation?
Yes. The REST API accepts image uploads, lets you specify the model and document type, and returns structured JSON with full text, confidence, and warnings. API keys are in Settings → API Keys. Webhooks are available for async result delivery.
How do I keep bulk OCR results consistent?
Run one representative image in compare mode, accept the best output, and save that model as a preference. All subsequent uploads start from that model automatically, so the same engine reads every image in your batch without manual selection each time.
Can I export bulk results as JSON for downstream processing?
Yes. Every accepted result — from the workbench or the API — exports as structured JSON that keeps full text, document type, confidence score, warnings, and detected tables. JSON is the right format for batch pipelines: metadata travels with the text so downstream systems can filter or flag results without reopening each file.
How are credits charged for bulk image to text?
Each OCR job charges that model's per-document credit cost. A single-model run on one image charges once; compare mode on the same image charges each selected model. For batch runs via the API, credit cost scales linearly with the number of jobs submitted. The free plan includes 100 credits every month to test your batch workflow before committing to a plan.