Questions about Ragflow

[xyzhu68]

• In your blog "The Rise and Evolution of RAG in 2024 A Year in Review" you talk about document intelligence models. DeepDoc belongs to the first generation. For the second generation: "The enterprise version of RAGFlow also uses this architecture for document processing". So is there an "enterprise version of RAGFlow"?

• The DeepDoc using models served on HuggingFace. Is DeepDoc running locally? Or will data be sent to cloud for parsing etc.?

• RAGFlow uses Elasticsearch by default. Is Elasticsearch your recommendation over Infinity (at version 0.19.1)?

Thank you!

[dosubot[bot]]

Yes, there is an enterprise (commercial) version of RAGFlow with features not available in the open-source version, such as RBAC, Department/Organization management, and advanced team management. The demo site (demo.ragflow.io) showcases these exclusive features, which are not present in the open-source release (source, more).

DeepDoc in RAGFlow runs locally by default. The open-source version uses ONNX models that you download from HuggingFace to your local environment—no data is sent to the cloud for inference unless you explicitly use an online SaaS LLM or embedding service. If your server can't access HuggingFace, you can manually download the model files and place them in the appropriate directory to ensure everything stays local (source, more, more).

Elasticsearch is the default and recommended backend for RAGFlow because it is stable, production-proven, and supports both full-text and vector search with advanced features. Infinity is an AI-native alternative with hybrid search and advanced ranking, but as of v0.19.1, it has known stability issues (such as crashes with large file counts). You can switch to Infinity by setting DOC_ENGINE=infinity, but Elasticsearch remains the safer choice for production right now (source, more).

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[ZhenhangTung]

So is there an "enterprise version of RAGFlow"?

Visit https://demo.ragflow.io/ and we're hosting our enterprise version there. If you're interested buying one, you can contact us.

The DeepDoc using models served on HuggingFace. Is DeepDoc running locally? Or will data be sent to cloud for parsing etc.?

Yes, DeepDoc is running locally.

RAGFlow uses Elasticsearch by default. Is Elasticsearch your recommendation over Infinity (at version 0.19.1)?

Pls choose Elasticsearch. We're refactoring Infinity.