'/%3e%3cpath%20d='M50%208%20L50%2050%20L15%2033%20Z'%20fill='%23FFFFFF'%20opacity='.18'/%3e%3cpath%20d='M50%208%20L85%2033%20L50%2050%20Z'%20fill='%23FFFFFF'%20opacity='.09'/%3e%3cpath%20d='M50%2050%20L85%2033%20L71.6%2078%20Z'%20fill='%23000000'%20opacity='.10'/%3e%3cpath%20d='M50%2050%20L28.4%2078%20L71.6%2078%20Z'%20fill='%23000000'%20opacity='.16'/%3e%3cpath%20d='M50%2050%20L15%2033%20L28.4%2078%20Z'%20fill='%23000000'%20opacity='.05'/%3e%3c/svg%3e)
Scattered
Saves live across bookmarks, socials, docs, notes, and screenshots.
Private beta now onboarding
Your saved internet, finally searchable.
Bookmarks, posts, articles, and notes — turned into a cited, navigable, agent-ready knowledge base.
X
Reddit
LinkedIn
URLs
Notes
Knowledge brain
AI Q&A
Smart feeds
Mind map
Cited search
Multi-source ingestionSelf-organising taxonomyFilesystem-backed knowledgeAgentic extraction pipelineAuto-indexed knowledge baseCited semantic searchChat, CLI, and MCP accessMulti-source ingestionSelf-organising taxonomyFilesystem-backed knowledgeAgentic extraction pipelineAuto-indexed knowledge baseCited semantic searchChat, CLI, and MCP accessMulti-source ingestionSelf-organising taxonomyFilesystem-backed knowledgeAgentic extraction pipelineAuto-indexed knowledge baseCited semantic searchChat, CLI, and MCP access
The problem
Saving is easy. Retrieval is the broken part.
Unstructured
Tools keep the link, lose the idea inside it.
Hard to reuse
You still have to remember where you saved it — and why.
Capabilities
Built around how you save and ask.
Connect every place you save from.
Bookmarks, articles, X, Reddit, LinkedIn, notes, docs — one corpus.
Feed
WEBMastering Database Connection Pooling
XThread on agent memory and durable summaries
LINKEDINHow we cut RAG latency by 40%
REDDITr/LocalLLaMA: scaling Postgres for embeddings
Table of contents
Database Engineering
Autonomous ML ResearchAI Agent MemoryTypes of LovePostgreSQL at ScaleDatabase Connection PoolingB-tree Indexes
Overview
Database engineering spans schema design, query planning, replication, and connection management.
PostgreSQL at Scale1 item
Connection Pooling1 item
Search history
what are the different types of loveDifferent types of love include the eight Greek-derived categories: Philia, Pragma, Storge, Eros…10h ago
How can we improve our RAG systems?Add persistent external memory, scale memory with the agent, and tighten retrieval grounding…2d ago
Navigate your knowledge like a book.
Saves snap into chapters, sections, and index paths automatically.
Ask your saved knowledge directly.
Grounded answers, always cited back to the source you saved.
// https://mcp.usenexus.in/mcp "mcpServers": { "nexus": { "url": "https://mcp.usenexus.in/mcp" "transport": "streamable-http" } }
Ask anything
AttachSearchReason
Agent-native by default.
Ships with a standards-based MCP endpoint for agentic tools and coding agents.
Agent access
Bring your saved knowledge into the agents you already use.
Nexus exposes a standards-based MCP server so agentic systems can retrieve your saved material with citations instead of guessing from chat history or stale project context.
ChatGPTClaude CodeCodexLangGraph agents
https://mcp.usenexus.in/mcpRemote Streamable HTTP, built for MCP-compatible clients.
Existing Nexus account required
The v1 flow uses hosted MCP auth and maps your verified email to your Nexus account.
Search, content, recent, status
Read-only tools retrieve cited answers and distilled saved content for the authenticated user.
Who it is for
Built for writers
who want their saved reading to turn into reusable context, not another unread queue.
Technical credibility
More than a bookmark folder.
CapabilityBookmark toolsNexus
Save linksYesYes
Manual foldersRequiredOptional
Keyword searchBasicSemantic and cited
Extract claims and conceptsNoYes
Book-like navigationNoChapters, sections, and index paths
Available to agents and toolsNoChat, CLI, and MCP
Self-updating knowledge mapNoYes
Early access
Request early access to Nexus.
Tell us where your saves live today.
FAQ
Frequently asked questions
Is Nexus just a bookmarking app?
No. Capture is only the entry point. Nexus is designed to extract, organize, and query the ideas inside your saved material.
How is it different from Readwise, Notion, or Obsidian?
Those tools are useful places to collect or write. Nexus focuses on turning scattered sources into a searchable, cited corpus organized like a living book, with chapters, sections, and index paths that reduce navigation overhead.
What sources will Nexus support?
The product direction includes bookmarks, URLs, articles, X posts, Reddit saves, LinkedIn posts, notes, docs, and local workspace knowledge through CLI and MCP integrations.
Will answers include citations?
Yes. Cited answers are central to the product. Nexus should show where an answer came from, not just produce detached summaries.
Why support chat, CLI, and MCP?
The interface should match the workflow. Some people will ask in the app, some will use a command line, and some will let an agent retrieve cited context through MCP.
When will I get access?
Early access will prioritize people with serious saved-knowledge workflows and teams testing source integrations, extraction quality, and agent-ready retrieval.