Early access

Build your
Open Brain.

Your knowledge should not live inside chat windows.

Sentiuma is a portable memory layer that turns your knowledge into a system every AI; Claude. ChatGPT. Cursor. can query. One brain.

See how it works

The flow

One brain.
Five moving parts.

01Markdown
where it starts
meeting-notes.md
---
project: speasy
people: dave
tags: product, audio
---
Dave suggested testing a shorter
audio digest format.
02Indexer
watches the folder
watching · ~/icloud/notes
roadmap-2026.md
ok
meeting-notes.md
hashing…
team-sync-oct.md
ok
dave-1-1.md
ok
03Knowledge
chunks, entities, links
speasy
dave
audio
product
notes
04MCP layer
the open door
mcp server
online
brain/search
brain/docs
brain/entities
brain/graph
05Any AI
reads the same brain
Cl
Claudeconnected
Ch
ChatGPTconnected
Cu
Cursorconnected

The product

Your knowledge,
finally visible.

The Open Brain dashboard surfaces everything your notes contain — documents, people, entities, and the connections between them.

localhost:3000/brain

Human interface

Your Open Brain

147

23

312

Vault

StatusConnected
Documents147
Entities312
Last indexed2 min ago

AI connections

Cl
Claude
not connected
Ch
ChatGPT
not connected
Cu
Cursor
not connected

Run sentiuma mcp start then add the server URL to your AI client.

MCP server

AI door
localhost:3001/mcplocal

The problem

Every AI you use
builds its own memory.

Claude remembers what you told Claude. ChatGPT remembers what you told ChatGPT. Those memories never connect. Over time, your knowledge spreads across models and none of them know the whole picture.

Claude
└ conversations
└ project notes
ChatGPT
└ research
└ strategy docs
Cursor
└ architecture
└ code decisions
// each tool knows a piece. none know the whole thing.

You repeat yourself

Every new tool, every new session — you re-explain projects, people, and past decisions. The real bottleneck in AI workflows is not prompts. It is memory.

Context resets constantly

Every chat window starts from zero. Ideas you already explored get re-asked. Nothing compounds. Each session is an island.

Your notes sit idle

You already have the knowledge. It lives in markdown files, in iCloud, in Obsidian. AI tools just cannot reach it.

Switching has a cost

Moving from Claude to ChatGPT to Cursor means losing context each time. The best model for the job should not require starting over.

The shift

Move memory
outside the models.

Instead of every AI building its own memory of you, create a brain they all read from. Your knowledge becomes infrastructure.

Before

  • Claude has its memory
  • ChatGPT has its memory
  • Cursor has its memory
  • They never connect
  • You repeat yourself

With Sentiuma

  • One brain, every AI
  • Write once, query anywhere
  • Knowledge compounds
  • Switch models freely
  • Your data stays yours

See it work

Ask anything.
Answered from your notes.

Connect Claude, ChatGPT, or Cursor through the MCP layer. When they answer, they read from your brain — not their training data.

Cl
Claude + Open Brain MCPconnected
What are the key product decisions we made last quarter?

Based on your notes, here's what I found from last quarter:

MCP integration takes priority — you decided to focus on MCP over the API rewrite, targeting January 2026.

Speasy Q2 launch confirmed — Dave signed off in the Oct 14 team sync. Budget was approved.

Roadmap locked through Q3 — indexer v2 (Q1), multi-vault (Q2), team workspaces (Q3).

Ask anything about your knowledge…
Sources3 matched

product-strategy-q4-2025.md

3 days ago

Decided to prioritise MCP integration over the API rewrite. Timeline: January 2026.

MCPAPI rewrite

team-sync-oct-14.md

6 days ago

Dave confirmed the API timeline. Speasy targeting Q2 launch. Budget approved.

DaveSpeasy

roadmap-2026.md

1 week ago

Q1: indexer v2 with streaming. Q2: multi-vault support. Q3: team workspaces.

indexer v2multi-vault

How it works

Simple infrastructure.
Serious leverage.

01

Write notes normally

Use Ulysses, Obsidian, VS Code — any markdown editor you already use. Your files remain the source of truth. Nothing changes about how you write.

meeting-notes.md
---
project: speasy
people: dave
tags: product, audio
---
Dave suggested we test a shorter
audio digest format for Speasy.
02

The indexer watches your folder

A lightweight macOS app monitors your iCloud markdown folder. When a file changes, it hashes the content, detects what is new, and sends updates automatically. Set and forget.

03

Notes become structured knowledge

The backend parses your content — chunking documents, generating embeddings, and extracting entities like people, projects, and ideas. Your notes become machine-readable without any extra work from you.

04

Every AI can query your brain

AI tools connect through an MCP server. They can search your notes by meaning, retrieve relevant documents, and explore relationships. Claude, ChatGPT, Cursor — they all read the same brain.

Surface what matters

Your brain notices
things you miss.

The signal engine scans your knowledge base weekly — surfacing unresolved decisions, threads going cold, and ideas worth pursuing. No extra work required.

Decision

API timeline not confirmed since Oct 14

Today

Dave signed off on January, but no follow-up has been logged. The decision may have shifted.

92%
·
2 sources
Stale

Speasy hasn't appeared in notes for 30 days

Yesterday

Previously a recurring topic across planning docs. May have stalled or been quietly deprioritised.

78%
·
1 source
Idea

MCP integration mentioned 5× in the last 2 weeks

2 days ago

Recurring across project notes, meeting logs, and the roadmap. Might be ready for a dedicated strategy document.

85%
·
3 sources
Gap

Dave's current project scope not documented

3 days ago

Dave appears in 14 documents but no note describes what he's actively working on.

71%
·
1 source

Signals are generated weekly from your indexed knowledge base.

Who it's for

Built for people who
work with ideas over time.

Sentiuma is not a note-taking app. It is the layer that makes your existing notes useful to AI. If you already keep knowledge in markdown and want AI to reason with it — this is for you.

Product designersEngineersFoundersResearchersWritersAnyone building knowledge