infiniflow/ragflow

Choose this fork if you need its custom UI/API/session behavior and are okay with carrying a materially diverged, behind-upstream codebase. Choose upstream if you want the latest fixes, broader language/search support, and lower maintenance burden.

Prefer this fork only if you need its custom behavior and can own the maintenance burden. If you want current RAGFlow features, fixes, and documentation, upstream is the safer choice.

Choose this fork only if its bespoke agent/tool and dataset workflows match your product needs and you are prepared to carry the upgrade burden. If you want active upstream parity, this fork looks too stale and too divergent.

Choose this fork only if you want a heavily modified RAGFlow base and are willing to own merge debt. If you want current upstream features, bug fixes, and easier upgrades, upstream is the safer choice.

Choose upstream unless you need the fork’s research-data and deployment customizations and are prepared to own a large maintenance burden. This fork looks like a specialized, stale branch rather than a drop-in replacement for current RAGFlow.

Prefer this fork only if its added admin and workflow changes are directly valuable and you are prepared to maintain a large downstream delta. If you want a stable, broadly current RAGFlow baseline, upstream is the safer choice.

Prefer this fork only if its extra workflow features match your product requirements and you are comfortable owning a large upstream delta. If you want broad upstream compatibility, multilingual coverage, and lower maintenance cost, upstream is the safer choice.

Choose this fork if you want a customized RAGFlow base and can support ongoing merge maintenance. Choose upstream if you want current fixes, lower risk, and less divergence.

Choose this fork only if its custom agent/API and model-support changes are the point. For most adopters, upstream is the safer default because it is much more active and this fork is far behind and heavily diverged.