Interactive notes
living ideas

No. 01 — Knowledge vault

A second brain for AI-native minds.

Drop in any file, link, document, or raw note. Noteshell turns it into a living, analyzed, interactable workspace — built around your thinking, not someone else's template.

Anything in, knowledge out.

Drop documents, screenshots, voice memos, links, raw data. Noteshell catches all of it — students with course folders, researchers with paper stacks, founders with scattered notes and meeting transcripts.

• Lossless capture — original sources kept
• OCR for screenshots, transcription for voice
• Live indexing across the entire vault
• Source citations on every reuse

Living notes, not static files.

A captured PDF doesn't sit there waiting to be read. It feeds into dynamic notes, queryable summaries, and reports that update as your collection grows. The vault keeps working after the capture.

Capture is the start. The notes keep working after.

No. 02 — Programmable workspace

Tell the AI what you need. Watch it appear.

Slide decks inside your notes. Journal dashboards with progress charts. A daily calorie tracker that reads your eating-plan note. Custom widgets pulling from Slack, Linear, Calendar — built when you ask.

Every artifact is yours to shape.

Notes, dashboards, components, canvases, slides — they're all the same primitive underneath. Rearrange, customize, build the shape your work actually needs. A writer wants a journal-progress dashboard; a student wants a slide deck nested in lecture notes; a runner wants a calorie tracker that reads today's meal plan. All buildable.

• Programmable dashboards from a prompt
• Living notes that compute, not just display
• Visualized graphs from any structured note
• Custom widgets connected via MCP

Connect to your stack.

Noteshell speaks MCP. Plug in Slack for notifications, Linear for the issue board, Google Calendar for the day's schedule, your own internal services. Then ask the AI to build the surface around it.

If you can describe it, you can build it.

No. 03 — Agent Canvas

Brainstorm across hundreds of notes at once.

Display many notes side-by-side on a single canvas. Connect them. Let an agent find the patterns, group the related, and produce summary components and graphs you can actually keep.

Built for complex synthesis.

Researchers summarizing across literature reviews. Founders synthesizing user interviews. Anyone whose work is "I have a lot of inputs and I need to see them connected." Agent Canvas makes the connection visible — and editable.

• Multi-note canvas view
• Auto-grouping by theme and content
• Generated summary components
• Graphs, charts, and reports inline

From inputs to insight, on one surface.

Drag a stack of notes onto the canvas. Ask the agent to summarize. The summary becomes a component you can wire into another note, or pipe into a chart. The whole synthesis lives in one place — re-runnable as new notes arrive.

The hidden patterns finally have a place to show up.

No. 04 — Second-brain at scale

A knowledge-first home — from one mind to millions of files.

Built for one user with a few hundred notes — and for organizations indexing millions. Same model, scaled. The thread doesn't break when the vault grows.

An encyclopedia for your team.

Connect Noteshell to your company's MCP servers. Index everything — docs, tickets, meetings, code. The team gets a complete, queryable second-brain that retrieves and analyzes at speed, with citations every team member can trust.

• Millions-of-files scale, no degradation
• MCP-native ingestion from any source
• Reliable retrieval with confidence reporting
• Org-wide knowledge graph, queryable by anyone

Publish what you build.

Found a note template that works? A dashboard that saves your team an hour a day? A canvas workflow that nails research synthesis? Publish it. Others adopt it, fork it, build on it. Your knowledge becomes a workflow others can run.

Your knowledge becomes a workflow others can run.