Fetching from the wire…
01
02
03
04
05
06
07
08
OSS2026-07-30 · source-backed
infiniflow/ragflow now lists Go as its primary language and describes itself as fusing RAG with agent capabilities to serve as a context layer. The repositioning from retrieval engine to context layer is the notable part, and it mirrors what smaller projects are doing from the opposite direction. RAG projects are converging on context management as the product, with retrieval as an implementation detail.
Each link below shares sources, entities, or timing with this story.
The August 28 release moves metadata filtering into the index instead of applying it post-retrieval, and opens rerank_candidates_count, knn top_k and num_candidates on the Retrieval API so the recall/precision tradeoff is tunable per call. It adds an Azure DevOps connector, Yo...
infiniflow/ragflow keeps momentum with agentic retrieval and context management features, positioning RAG as an intelligent agent subsystem rather than a static pipeline. The highest-starred dedicated RAG engine on GitHub and the open-source benchmark to evaluate against.
1. Flip your multi-model pipeline to review-then-generate. Instead of using a reasoning model to plan before code generation, let the specialist generate freely and use reasoning tokens for review. Paper shows 90.2% pass@1 vs 87.2% for the planning pattern. Source 2. Audit you...
Ryan Dahl announced celld on August 5. It's a daemon built from V8, S3, SQLite, LTX and Tokio that runs the exact Cloudflare Workers and Durable Objects JavaScript APIs and configuration on hardware you own. The architecture is the interesting part, not the API compatibility....
v0.2.11, August 25, makes Flash the local default for its vectorless RAG: the tree comes from layout statistics, LLMs write only node summaries, tree expansion proposes a wave of nodes concurrently rather than one round-trip at a time, and embedded PDF bookmarks get consumed w...
Andrej Karpathy published a GitHub Gist describing a pattern I've been using without having a name for it. He calls it the "idea file" pattern, and it flips the RAG orthodoxy on its head. The concept: instead of retrieving documents at query time (the standard RAG approach), u...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.