Hi dots.ocr
I hope this finds you well.
I am the developer of Nexo ♾️, an adaptive cognitive ecosystem built on Python and Obsidian, designed for large-scale knowledge management and "Cognitive Continuity" (AAAK Architecture). I’ve been following the development of Dots-OCR and I am very impressed with its layout-aware parsing capabilities.
I am writing to let you know that I plan to integrate your engine as a core normalization layer in my v5.1 Pipeline. My application follows a strict "Zero Token Cost / Free to Use" philosophy, and your MIT-licensed work fits perfectly with this vision.
How I plan to use it:
Media Ingestion: Converting PDFs and images into structured Markdown before they enter my "Triple Filter" (Hash -> Similarity -> Heuristics).
Idle Processing: Using your engine during "Idle Mode" to extract "DNA" from high-volume documents (>130k words) and store it in a non-volatile cache within Obsidian.
MemAtom Generation: Providing high-quality structured text to my AI enrichers to ensure technical accuracy in metadata.
I’ll be performing some fine-tuning to ensure the output alignment with my system's requirements. If you have any specific recommendations regarding resource management (CPU/GPU) for background processing or tips for optimal Markdown output structure, I’d love to hear them!
Thank you for your amazing contribution to the OCR space. I will properly credit your work in my documentation and license notices.
Best regards,
Ccom5 (Grabriel)
Developer of Nexo ♾️
https://github.com/Ccom5/Nexo-
Hi dots.ocr
I hope this finds you well.
I am the developer of Nexo ♾️, an adaptive cognitive ecosystem built on Python and Obsidian, designed for large-scale knowledge management and "Cognitive Continuity" (AAAK Architecture). I’ve been following the development of Dots-OCR and I am very impressed with its layout-aware parsing capabilities.
I am writing to let you know that I plan to integrate your engine as a core normalization layer in my v5.1 Pipeline. My application follows a strict "Zero Token Cost / Free to Use" philosophy, and your MIT-licensed work fits perfectly with this vision.
How I plan to use it:
Media Ingestion: Converting PDFs and images into structured Markdown before they enter my "Triple Filter" (Hash -> Similarity -> Heuristics).
Idle Processing: Using your engine during "Idle Mode" to extract "DNA" from high-volume documents (>130k words) and store it in a non-volatile cache within Obsidian.
MemAtom Generation: Providing high-quality structured text to my AI enrichers to ensure technical accuracy in metadata.
I’ll be performing some fine-tuning to ensure the output alignment with my system's requirements. If you have any specific recommendations regarding resource management (CPU/GPU) for background processing or tips for optimal Markdown output structure, I’d love to hear them!
Thank you for your amazing contribution to the OCR space. I will properly credit your work in my documentation and license notices.
Best regards,
Ccom5 (Grabriel)
Developer of Nexo ♾️
https://github.com/Ccom5/Nexo-