Knowledge Base / RAG
SaaS, open source and skills for knowledge base / rag work — with the limits of each written down.Before you pick
What decides it
Your corpus is mostly near-duplicates — drafts, revisions, superseded policies. Similarity alone can't rank them; ask what breaks the tie: recency, ownership, version.
What to avoid
They respect permissions but end obscurity. Sharing links set too broadly years ago become one question away, and the cleanup lands after go-live.
What is changing
Retrieval is becoming a planned loop: the model splits one question into several searches, reranks each, then merges. Slower and costlier per answer.
SaaS
Hosted products you sign up for.Leading
Established picks, at the top of today's aggregate.
- 1GleanEnterprise workplace search and RAG across many connected tools; strongest for broad cross-app knowledge retrieval.
- 2GuruAI knowledge base that surfaces verified answers inside your workflow with source trust/verification built in.
- 3Notion AITeams already in Notion who want RAG Q&A over their own docs and wiki without standing up separate infra.
- 4Confluence + Atlassian IntelligenceAtlassian shops that want AI search and answers layered onto existing Confluence documentation.
- 5DashworksAI assistant that answers questions across all connected company apps without pre-indexing everything.
Emerging
Newer challengers, surfaced by two or more sources.
- 1Link
- 2HyperengageMemory layer that fuses CRM, product usage, support and billing records into one graph per customer account for post-sales AI agents.
- 3WetrocloudAPIs for building RAG applications — document extraction, markdown conversion, retrieval endpoints — plus a gateway routing requests across hundreds of hosted models.
- 4vecr.ioIndexes a GitHub repo, website or documents and serves the result as an MCP search endpoint, with no embedding infrastructure to run.
Open source
Repositories you run yourself.Leading
- 1LangChainThe de-facto framework for building RAG and knowledge-base apps — chains, agents, and connectors to every model and vector store. The category's recognized leader (~142k stars).
- 2DifyOpen-source LLM app platform with a built-in knowledge base, document ingestion, and RAG pipeline behind a full UI; the highest-starred option here (~150k stars).
- 3RAGFlowDedicated open-source RAG engine with deep document parsing, knowledge-base management, and a zero-code web UI (~86k stars).
- 4AnythingLLMAll-in-one desktop/self-hosted app that turns your documents into a private chat knowledge base out of the box (~64k stars).
- 5LlamaIndexThe data framework for indexing and querying your private content; the reference toolkit for document-QA knowledge bases (~51k stars).
Emerging
- 1Zleap-AI/SAGLocal-first knowledge base desktop app whose retrieval indexes extracted events and entities, building graph links at query time rather than prebuilt vector-plus-graph pipelines.
- 2ConardLi/rag-skillAgent skill that answers questions over local PDF, Excel and Markdown files using layered directory indexes and grep instead of embeddings.
- 3boluo2077/deep-ragSelf-hosted FastAPI and React app where the model calls tools to walk whole documents, handling multi-hop, negation and aggregation questions vector search misses.
- 4phoenix-zhou/logistics-industry-RAGReference implementation of an offline document question-answering pipeline using FAISS and locally run Qwen or ChatGLM models, demonstrated on logistics documents.
- 5upupmake/general-rag-systemMulti-tenant enterprise knowledge base with Vue, Spring Boot and FastAPI services, agentic retrieval loops, workspace permissions and MCP access for outside agents.
Skills
Skills you drop into a coding agent.Leading
- 1obsidian-skills (kepano)Official Obsidian Agent Skills that teach an agent to read/write your Markdown vault, Bases, and Canvas as a knowledge base (~43k stars).
- 2claude-obsidian (AgriciDaniel)Self-organizing second-brain skill set: drop any source and Claude reads, links, and files it into one connected Markdown knowledge graph (~9.9k stars).
- 3obsidian-second-brain (eugeniughelbur)Cross-CLI skill turning a vault into an AI-first knowledge base with hybrid semantic search and self-maintaining notes (~3.5k stars).
Emerging
- 1topoteretes/cogneeOpen-source Python library and hosted service that turns documents, chats and app data into a queryable knowledge graph agents recall with citations.
- 2alibaba/zvecIn-process vector database from Alibaba that embeds in your app instead of running as a server, with dense, sparse and hybrid search.
- 3chappyasel/meta-kbCompiled wiki of 155 articles on agent memory, context engineering and multi-agent design, rebuilt from cited sources by an automated verification pipeline.
- 4bsisduck/google-search-ads-analytics-docsOfficial Google docs for Search, Search Console, Ads and GA4 mirrored as local Markdown, with skills for cited retrieval and SEO audits.
- 5q8886b/loom-ai-knowledge-network