بصير Baseer
A vision-language model built for Arabic document structure. Scans, PDFs, images, and tables become structured data, contracts, financial statements, licenses, shipping and customs paperwork, without a translation layer getting in the way.
- SOTA accuracy on Arabic document benchmarks (Arabic handwritten text recognition, WER 0.24 / CER 0.08)
- Live self-serve trial today at baseerocr.com
- Pay-as-you-go API, no fixed monthly fee, volume-tiered from $5.00 down to $3.50 per 1,000 pages
- Deployed for contract extraction, audit documentation, and port/customs automation at government and enterprise scale