🚧 Work in Progress: This project is under active development.

A Model Context Protocol (MCP) server for org-mode knowledge management. Provides search, content access, and note linking capabilities for your org-mode files through the MCP protocol.

• org:// — List all org-mode files in configured directories
• org://{file} — Access raw content of {file}
• org-outline://{file} — Get hierarchical structure of {file} as JSON
• org-heading://{file}#{heading} — Access specific headings by path
• org-id://{id} — Find content by org-mode ID properties
• org-agenda:// — List all agenda items and tasks
• org-agenda://today — Today's scheduled agenda items
• org-agenda://week — This week's scheduled agenda items
• org-agenda://day/{YYYY-MM-DD} — Agenda for a specific day
• org-agenda://week/{N} — Agenda for week number N
• org-agenda://month/{N} — Agenda for month number N
• org-agenda://query/from/{YYYY-MM-DD}/to/{YYYY-MM-DD} — Custom date range

• org-file-list — List all org files in configured directories
• org-search — Full-text fuzzy search across all org files
• org-agenda — Query agenda items with filtering by dates, states, tags, and priorities
• org-capture — Append a heading to an org file. Supports TODO state, priority, tags, body, SCHEDULED/DEADLINE/CLOSED timestamps (with repeater and warning suffixes), property drawer entries, and datetree expansion.
• org-update-todo — Update an existing heading in place. Set or clear todo_state, priority, tags, planning timestamps, heading title, body text, and property drawer entries (per-key upsert/remove).

This repository ships skills for Claude Code, Codex, Cursor, and OpenCode that guide agents to use the MCP server effectively.

claude plugin install github:szaffarano/org-mcp-server

Or add to .claude/settings.json:

{ "plugins": ["github:szaffarano/org-mcp-server"] }

Add to your plugin configuration using .codex-plugin/plugin.json or .cursor-plugin/plugin.json from this repository.

{ "plugin": ["org-mcp-server@git+https://github.com/szaffarano/org-mcp-server.git"] }

See .opencode/INSTALL.md for full setup instructions.

Download the latest pre-built binaries from GitHub Releases:

curl -LO https://github.com/szaffarano/org-mcp-server/releases/latest/download/org-mcp-server-x86_64-unknown-linux-gnu.tar.gz
tar xzf org-mcp-server-x86_64-unknown-linux-gnu.tar.gz
sudo mv org-mcp-server /usr/local/bin/

Binaries are available for Linux, macOS, and Windows. Check the releases page for all targets.

cargo install org-mcp-server --locked
cargo install org-cli --locked

nix run github:szaffarano/org-mcp-server       # run directly
nix profile install github:szaffarano/org-mcp-server  # install to profile
nix develop github:szaffarano/org-mcp-server   # development shell

git clone https://github.com/szaffarano/org-mcp-server
cd org-mcp-server
cargo build --release

Add to your AI agent configuration (e.g., ~/.claude.json, opencode.json):

{
  "mcpServers": {
    "org-mode": {
      "command": "/path/to/org-mcp-server",
      "args": [],
      "env": {
        "ORG_ORG__ORG_DIRECTORY": "/path/to/your/org/files"
      }
    }
  }
}

Or via Nix:

{
  "mcpServers": {
    "org-mode": {
      "command": "nix",
      "args": ["run", "github:szaffarano/org-mcp-server"]
    }
  }
}

Config file: ~/.config/org-mcp/config.toml (or $XDG_CONFIG_HOME/org-mcp/config.toml).

[org]
org_directory = "~/org/"
org_default_notes_file = "notes.org"
org_agenda_files = ["agenda.org", "projects.org"]
org_todo_keywords = ["TODO", "|", "DONE"]
# Auto-prepend :CREATED: property on capture (default: true)
org_auto_created_property = true
# Auto-stamp CLOSED on done transitions (default: true)
org_auto_closed_timestamp = true

[logging]
level = "info"
file = "~/.local/share/org-mcp-server/logs/server.log"

[cli]
default_format = "plain"  # plain | json

Configuration is resolved in order: CLI flags → environment variables (ORG_ prefix, __ separator, e.g. ORG_ORG__ORG_DIRECTORY) → config file → defaults.

org-cli config init   # create default config file
org-cli config show   # display current resolved config
org-cli config path   # show config file location

org-cli mirrors the MCP server's capabilities for scripting and testing. Run org-cli --help or org-cli <subcommand> --help for full usage.

Quick examples:

# Search across all org files
org-cli search "project planning"

# Agenda
org-cli agenda today
org-cli agenda list --states TODO,IN_PROGRESS --tags work

# Capture a TODO with planning
org-cli capture "Fix login bug" --todo-state TODO --priority A \
    --scheduled "2026-05-15" --deadline "2026-05-20 -3d"

# Capture under a datetree
org-cli capture "Standup notes" --file journal.org --target-heading Logs --datetree

# Update an existing heading
org-cli update-todo --id abc123 --todo-state DONE
org-cli update-todo --file projects.org --heading-path "Work/Task" \
    --title "Renamed task" --property "EFFORT=2h"

Timestamp grammar for --scheduled, --deadline, --closed: YYYY-MM-DD [HH:MM] [+N{h|d|w|m|y} | ++N{u} | .+N{u}] [-N{u}]

Multi-crate Rust workspace:

• org-core — Business logic and org-mode parsing
• org-mcp-server — MCP protocol implementation
• org-cli — CLI interface for testing and direct usage

Built with orgize (org-mode parsing), rmcp (MCP protocol), tokio (async runtime), and nucleo-matcher (fuzzy search).

cargo test                                              # run all tests
cargo test -p org-core                                  # test specific crate
cargo fmt && cargo clippy --all-targets --all-features  # format and lint
cargo run --bin org-cli -- --help                       # run CLI

• File discovery and listing
• Basic content access via MCP resources
• Org-mode parsing with orgize
• ID-based element lookup
• CLI tool for testing
• Full-text search across org files

• Configuration file support with TOML format
• Environment variable configuration
• Unified CLI interface with global configuration
• Tag-based filtering and querying
• Agenda-related Functionality
• Link following and backlink discovery (org-roam support)
• Metadata caching for performance

• Content creation via org-capture (heading + planning + properties + datetree)
• Content modification: TODO state, priority, tags, and planning updates via org-update-todo / org-cli update-todo
• Content modification: property drawer updates (upsert/remove individual keys)
• Content modification: CLOCK / LOGBOOK entries
• org-promote / org-demote — relative heading level change
• org-clock — clock in / clock out / cancel clock
• org-refile — move heading to different file or location
• org-archive — archive heading to archive file or toggle ARCHIVE tag
• Media file reference handling
• Integration with org-roam databases
• Real-time file watching and updates
• Advanced query language

MIT License - see LICENSE file for details.