Send us a note
Use the form below to reach the maintainers of DeepSeekOCR.org. We triage questions daily and match you with engineers or community moderators who can help.
Direct channels
Support Inbox
support@deepseekocr.org
Priority assistance for integrations and bug reports.GitHub Discussions
github.com/deepseek-ai/DeepSeek-OCR/discussions
Roadmaps, RFCs, and long-form technical dives.Community Hub
community.deepseekocr.org
Meet contributors across language and industry verticals.Frequently asked questions
Yes. The model weights are released under a permissive license. We maintain guidance on attribution and usage boundaries in our Terms of Use. Your deployment must still respect local data regulations.
We coordinate with a network of service providers who specialize in scaling DeepSeek OCR. Share your requirements and we will connect you with partners or internal advisors capable of helping.
Open an issue or pull request in the GitHub repository. If you are proposing a large feature like tokenizer changes or dataset releases, start with a discussion thread so the community can review design considerations.
No, DeepSeekOCR.org does not run tracking scripts. Basic server logs are retained for 30 days for security audits and then purged.
Our Commitment
We are determined to make open multimodal OCR usable in production settings. Every message receives a thoughtful, actionable response from someone who has shipped DeepSeek OCR in the field.