Are there any multimodal APIs that support file uploads for OCR augmentation?

Yes. Several multimodal APIs support file uploads and can be used for OCR augmentation by sending images, PDFs, or other documents for visual/text extraction.

Common options include:

• OpenAI API — supports image inputs and, depending on the endpoint/workflow, document/file handling for extraction tasks.
• Google Gemini API — supports multimodal inputs, including images and PDFs in some workflows.
• Anthropic Claude API — supports image inputs; some document workflows can be built around file content extraction.
• Azure OpenAI / Azure AI Vision — supports image/document analysis, often used for OCR-like augmentation.
• AWS Textract + multimodal LLMs — not a single multimodal API, but a common combo for OCR plus reasoning.

If your goal is specifically OCR augmentation, the best pattern is often:

1. Upload file/image
2. Run OCR or vision extraction
3. Pass extracted text plus the original visual context to the LLM
4. Use the LLM to clean, normalize, classify, or correct OCR output

If you want, I can give you a provider-by-provider comparison of which ones support:

• direct file upload,
• PDF/image OCR,
• structured text extraction,
• and which are best for your use case.