RAGFlow v0.25.6 Adds AI Browser Agent and Upgrades RAPTOR with Ψ-RAG

By FintechExtra
27 May 2026

On May 26, 2026, the team behind RAGFlow released version 0.25.6, packing two notable upgrades: a new AI-powered browser agent and a revamped RAPTOR indexing method. These updates aren't just incremental—they nudge the platform toward more autonomous document processing and smarter semantic retrieval.

What Changed

First, the Agent module now includes a Browser component. This isn't your average web scraper. It's designed to let AI agents navigate and interact with web pages autonomously. Think of it as giving your AI a pair of digital hands to browse, click, and extract information from live sites. That's a big leap from static document ingestion.

Second, the RAG pipeline's RAPTOR construction gets a major upgrade with the introduction of AHC mode—dubbed Ψ-RAG (Psi-RAG). Previously, RAPTOR built hierarchical summaries from document chunks. Now, with AHC (Agglomerative Hierarchical Clustering), it expands semantics from the document level to the dataset level. The result? Index construction is significantly faster. And it doesn't stop there—the new approach outperforms the old RAPTOR in retrieval quality.

Why It Matters

For anyone building AI applications that need to access real-time web data, the Browser component is a game changer. It eliminates the need for custom scraping scripts, letting your AI fetch and process live information on the fly. Imagine a customer support bot that can check a product page for updates before answering. That's the kind of autonomy we're talking about.

The Ψ-RAG upgrade, meanwhile, tackles a core problem in retrieval-augmented generation: semantic breadth. By moving from document-level to dataset-level semantics, the system understands broader contexts. And because it's faster, you can index larger corpora without sacrificing performance. For enterprise deployments, that's a direct boost to throughput and accuracy.

Both features underscore a trend: RAGFlow is pushing toward more intelligent, autonomous RAG pipelines. The Browser component brings agency; Ψ-RAG brings scalability. Together, they make the platform a stronger contender for production AI workloads.

Official Source: https://github.com/infiniflow/ragflow/releases/tag/v0.25.6