What Blueprint Nexus is
Blueprint Nexus is one workspace for notes, documents, tasks, goals, and calendar events. Each item you create links to the others as you write, so context is something you navigate rather than something you hunt for inside nested folders. Nothing has to be filed in exactly one place, and nothing gets lost because it was filed in the wrong one.
The name is the idea. Every entry becomes a node in a knowledge graph, every reference becomes an edge, and that graph is a primary way to read your own work rather than a report generated at the end. blueSC at bluesc.org is shorthand for the same product; it is independent and is not affiliated with Bluesky, the AT Protocol, or any health insurance organization.
- One workspace, not five tabs — Notes, documents, tasks, reminders, and calendar events share a single model, so a task and the note that explains it are never out of sync.
- Edit like plain text — Documents stay readable Markdown, exportable at any time, with collaborative editing and revision history layered on top.
- Start free — The Free tier is permanent and includes note editing, task boards, living documents, and the visual knowledge graph.
How the knowledge graph connects your work
As you capture content, agents read it and propose connections: a research note links to the task it unblocks, a decision record links to the project it belongs to, a journal entry links to the goal behind it. You can accept, reject, or relabel any connection, so the graph reflects your judgment rather than an opaque automated verdict.
Nodes and edges are stored in PostgreSQL and traversed with recursive CTEs, which is what makes movement feel immediate: opening a node shows its neighbourhood instead of a paginated table. The same structure renders as a clear 2D map for orientation and in 3D when you want to walk through the shape of a large body of work.
Planning, writing, and tracking in one place
Planning surfaces read from the same graph as your writing. The daily planner, kanban board, goals page, reminders, and calendar are views over the same nodes and edges, so moving a task updates the graph, and every document that links to it sees the change immediately.
Writing is a first-class surface too. Living specs and documents support collaborative editing, inline comments, and revision history, and an agent can draft, restructure, or summarise a document while keeping a visible citation trail back to the nodes and sources it used.
- Capture first, organize later — Write now and let structure emerge from the content, instead of choosing a folder before you know what you are writing.
- Views, not copies — Board, agenda, calendar, and graph are projections of one underlying workspace, so there is no second copy to reconcile.
Governed AI agents that work on your graph
Agents here are governed rather than open-ended. A tick-based execution loop runs a strict two-pass model: Pass A researches read-only and produces a proposal, and Pass B executes only the approved steps. Every tool call, risky action, and routing decision is written to an audit trail you can read line by line.
The fleet defaults to DeepSeek V4.1 Flash, routed through OpenRouter's CoreWeave FP8 endpoint so reasoning stays fast and inexpensive, and the same default applies across every agent workflow. Pro users can bring their own keys for OpenAI, Anthropic, or any OpenAI-compatible endpoint and keep control of both cost and data path.
- Risk-gated tools — Actions are classified before they run, and anything destructive waits for an explicit approval step rather than improvising.
- Open to other clients — The same workspace is exposed over the Model Context Protocol, so Cursor, Claude, ChatGPT actions, and custom agents read and write the graph you already have.
Who Blueprint Nexus is for
It is used by students assembling a term of study material, researchers tracking sources and open questions, writers holding chapters and story threads together, product and engineering teams documenting systems beside the tasks that change them, and anyone who wants a second brain that is still legible years later.
You do not need to be technical to use it. The graph builds itself in the background while you write, and agent features stay out of the way until you ask for them, while developers who want more can connect repositories, terminals, and MCP clients to the very same workspace.
How to get started
Create a free account and a starter workspace is provisioned for you, with nothing to configure before your first note. Import the documents you already have, switch on the integrations for tools you already use, and invite the first collaborator when the workspace is ready to share.
If you would rather look before signing in, the sandbox runs the full interface with example content, and the help center walks through the knowledge graph, the agent fleet, and model settings one screen at a time.
