Ai

RAGFlow v0.25.0 Sharpens GraphRAG, SSO, and Service Layer Architecture

RAGFlow v0.25.0 Sharpens GraphRAG, SSO, and Service Layer Architecture

RAGFlow v0.25.0 is a focused maintenance and platform refinement release that improves reliability across document ingestion, authentication, model configuration, and GraphRAG workflows. The update also continues backend modularization by splitting memory and message APIs into clearer gateway and service layers, making this version more important for teams tracking architectural progress as well as bug fixes.

What Changed

The most notable functional fixes in v0.25.0 address several areas that can directly affect production deployments. RAGFlow fixed image file uploads, corrected the Gemini embedding model name in llm_factories.json, made time utility tests timezone-independent, and resolved GraphRAG extraction issues. The release also fixes SSO auth token persistence on the root route loader, which should reduce authentication friction for users relying on single sign-on flows.

On the backend architecture side, the project refactored both the memory API and message APIs by splitting them into gateway and service layers. While this may not immediately change end-user behavior, it signals continued internal cleanup that should make the codebase easier to maintain, test, and extend in future releases.

Beyond code changes, the release includes sandbox reference documentation updates, refreshed v0.24.0 release notes, an Aliyun repository change, test file reorganization, and a bundled bug-fix pull request that addresses additional issues under the umbrella of general stability improvements.

Why It Matters

For AI teams using RAGFlow in real retrieval-augmented generation pipelines, this release matters because it improves several practical failure points rather than introducing flashy but risky new features. Better GraphRAG extraction and corrected embedding configuration reduce the chance of bad retrieval behavior, while upload and SSO fixes improve day-to-day usability for operators and enterprise users.

The gateway-and-service-layer refactors are also worth watching. They suggest the maintainers are preparing the platform for cleaner service boundaries, which typically supports better scalability, easier integrations, and faster future iteration. In that sense, v0.25.0 looks less like a headline feature release and more like a strategically important hardening update for organizations building on top of RAGFlow.

Official Source: https://github.com/infiniflow/ragflow/releases/tag/v0.25.0