Back to Morion

MCP Integration

33 tools your AI can call.Read notes. Claim tasks. You control the access.

Morion is an MCP server. Connect any AI assistant — it can search notes, pull context, claim tasks from your kanban, and move cards through statuses. All from a local SQLite file on your machine. No cloud, no API keys, no network.

What is MCP?

Model Context Protocol (MCP) is an open standard that lets AI assistants call tools on external servers. Instead of pasting context into every chat, your assistant connects to Morion and pulls what it needs on its own.

Morion

SQLite on your disk

MCP (stdio)

Claude

Codex

Gemini

Cursor

Cline

Zed

Any MCP client

Same protocol, same tools

Mo + MCP — they compose

MCP is how external AI clients reach Morion. Mo is the assistant that lives inside it.

They're not redundant — they're a team. Mo runs the workspace: organises your data, finds anything fast, drafts updates, watches your projects, asks before anything risky. Your MCP-connected agents (Claude Code, Codex, Gemini Antigravity, Cursor, Cline, Zed) do the actual work — write the code, the docs, the designs.

Same notebook, same kanban, same audit log. Every action is logged with the actor — "claude-code", "codex", "cursor", "Mo", "user" — so you always know who did what.

On Mo Managed (coming soon), Mo upgrades to a Context Broker: when Claude Code or Codex starts on a card, it asks Morion for the work packet — task, spec, decisions, recent activity — and Mo packages it from your real data. No more "read CLAUDE.md from scratch every session."

Quick Setup

Add one entry to your client's config file. That's it — no daemon, no API key, no network.

Claude Desktop claude_desktop_config.json

{ "mcpServers": { "morion": { "command": "morion", "args": ["mcp"] } } }

Cursor / Cline / Zed mcp.json

{ "morion": { "command": "morion", "args": ["mcp"] } }

Same shape works for any MCP client that supports stdio transport.

Available Tools

33 tools — all free

Every tool is available in the free tier. No paywalls on functionality — Pro adds permissions, not tools.

Notes (10 tools)

notes_search	Hybrid keyword + semantic search across all notes
notes_list	List notes with folder/tag/pinned filters and pagination
notes_get	Fetch a single note by ID with full body and tags
notes_create	Create a new note with optional folder, tags, and pinned state
notes_update	Partially update a note — only provided fields change
notes_delete	Soft-delete a note (moves to trash, 7-day retention)
notes_append	Append text to an existing note without rewriting it
notes_duplicate	Clone a note with a fresh ID and timestamps
notes_move	Move a note to a different folder
notes_recent	Last N notes by updated_at — "what was I working on?"

Folders (8 tools)

folders_list	List all folders ordered by position
folders_create	Create a new folder with optional parent
folders_rename	Rename a folder
folders_delete	Delete a folder — notes inside become unfiled
folders_duplicate	Clone a folder with all its notes
folders_move	Move a folder up or down in the sidebar
folders_reorder	Reorder all folders by passing an ordered ID list
folders_set_view_mode	Toggle a folder between list and kanban view

Tasks (4 tools)

tasks_list	Read tasks in a kanban folder, filter by status and date
tasks_move	Move a card between columns or reorder within a column
tasks_claim	Atomically claim a todo task — moves to doing, no races between agents
tasks_history	Full status-transition history for a task — who moved what, when

Six statuses: note · backlog · todo · doing · review · done. A task is just a note with a status — same search, same folders, same MCP.

Comments & activity (4 tools)

notes_list_activity	Paginated activity feed — comments + audit entries unified, cursor-based
notes_add_comment	Post a comment on a note (with optional 1-level reply)
notes_update_comment	Edit a previous comment (actor must match)
notes_delete_comment	Remove a comment — actor-match enforced, audit-logged

Discuss a note or task with your agents — three speakers (you, the LLM body author, anyone commenting). Audit and comments stream into one timeline.

Attachments (2 tools)

notes_list_attachments	Metadata-only listing — ~30 tokens per attachment, lazy-loaded
notes_get_attachment	Returns MCP ImageContent — vision-tokenized, ~1-5k tokens vs ~250k/MB for base64

Paste a screenshot in your notebook. Your AI client retrieves it as a real image, not a base64 wall.

Tags (4 tools)

tags_list	List all tags
tags_create	Create a tag with optional color
tags_update	Rename or recolor a tag
tags_delete	Delete a tag — notes keep their content

Audit (1 tool)

audit_recent	Last N audit log entries — see what each AI client changed

Every MCP write is logged with the client name. You always know what changed and who did it.

Morion Pro

Control what your agents see and do.

Your work notes, your journal, and your production deploy button shouldn't all have the same agent access. Pro draws the lines — folder visibility, read-only vs read-write, and status transitions allowed per agent.

Folder-level visibility

Choose which folders your AI can see. Private folders are completely invisible to MCP — they don't show up in search results, tool calls, or folder listings. Your journal stays yours.

MCP Folder Access

Work visible

Projects visible

Journal hidden

Health hidden

Read-only or read-write per folder

Some folders your AI should read but never modify. Meeting notes, reference docs, project specs — let it search and quote, but not rewrite. Other folders you want it to write to freely. Set it per folder.

Access Level

Work read + write

Projects read + write

Meeting Notes read only

Reference read only

Per-note overrides

A folder rule covers most cases, but a single sensitive note can opt out independently. Three flags per note (visible / update / delete) override the parent folder's defaults.

Inside Work folder

Q2 OKRs · draft visible · editable

Hiring plan visible · editable

Acquisition memo (sealed) invisible to MCP

Folder is visible by default. One note overrides it.

From $4/mo

$4/mo billed annually ($48/yr) — or $5.99/mo monthly

Per-folder + per-note MCP access permissions, unlimited kanban boards, and Mo running on your own LLM key.

Upgrade to Pro

MCP Permissions Compared

Most tools give AI full access or no access. Morion Pro is the only one with per-folder, per-client granularity.

Morion Pro	Obsidian	Notion	Mem	Bear
Folder-level visibility	Yes	~ plugin	No	No	No
Per-folder visible / create / update / delete flags	Yes	No	No	No	No
Per-note access overrides	Yes	No	No	No	No
Audit log of AI writes	Yes	No	No	Yes	No

~ partial / community plugin based on public documentation as of April 2026

Ready to try it?

macOS app is live. Download free, connect your AI client, and Morion shows up in your tools list.

Download for Mac See Pro