AllOurThings MCP Server
Enables cataloging and managing personal inventory (items, attachments) through natural language, allowing users to add, search, update, and retrieve item details and attachments via MCP tools.
README
AllOurThings
Your things, understood by AI.
AllOurThings is an inventory system that works the way you do. Catalog anything you like from your home appliances to your Pokémon cards — then ask plain-English questions and get instant answers.
Website: allourthings.io
Packages
| Package | npm | Description |
|---|---|---|
packages/mcp-server |
@allourthings/mcp-server |
MCP server — connects your inventory to Claude Desktop and other MCP clients |
packages/cli |
@allourthings/cli |
CLI — manage your inventory from the terminal |
Quick start
Desktop only. Requires macOS, Windows, or Linux with Claude Desktop or another MCP-compatible client.
1. Add to Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"allourthings": {
"command": "npx",
"args": ["-y", "@allourthings/mcp-server", "--data-dir", "~/Documents/AllOurThings"]
}
}
}
Restart Claude Desktop. Your inventory vault will be created automatically on first use.
2. Start asking questions
- "Add my Bosch washing machine, bought from John Lewis for £649 in January 2024 with a 2-year warranty"
- "What appliances do I own?"
- "What's in the kitchen?"
- "When does my TV warranty expire?"
- "Search for anything Samsung"
- "How much have I spent on electronics?"
How it works
The MCP server exposes your inventory to any MCP-compatible AI client via 10 tools:
| Tool | Description |
|---|---|
add_item |
Add a new item to your inventory |
get_item |
Retrieve an item by ID or name |
list_items |
List all items, optionally filtered by category, location, or tags |
update_item |
Update fields on an existing item |
delete_item |
Delete an item by ID |
search_items |
Full-text search across all item fields |
add_attachment |
Attach a file (manual, receipt, photo, warranty) to an item |
get_attachment |
Retrieve an attachment as base64 |
delete_attachment |
Remove an attachment from an item |
attach_from_url |
Download a file from a URL and attach it to an item |
Data
Vault structure
Your inventory lives in a vault — a plain directory on your filesystem. Each item gets its own folder:
~/Documents/AllOurThings/
items/
dyson-v15-detect-a1b2c3d4/
item.json
manual.pdf
receipt.jpg
samsung-65-qled-tv-b5c6d7e8/
item.json
warranty.pdf
Attachments (manuals, receipts, photos) sit alongside the item JSON. You can browse and edit the vault directly in Finder or File Explorer.
Item schema
Every item has required fields (id, name, created_at, updated_at) and well-known optional fields:
category brand model purchase_date purchase_price currency warranty_expires retailer location features notes tags attachments
The attachments field links PDFs and images stored in the item's folder:
{
"attachments": [
{ "filename": "manual.pdf", "type": "manual" },
{ "filename": "receipt.jpg", "type": "receipt" },
{ "filename": "photo.jpg", "type": "photo" }
]
}
You can also add any custom fields you like — they are preserved as-is.
CLI
A standalone terminal tool for power users and scripting. Works on macOS, Windows, and Linux. No AI client required.
# Run without installing
npx @allourthings/cli list
# Or install globally
npm install -g @allourthings/cli
Commands
allourthings search <query> # full-text search across all fields
allourthings list [--category <c>] [-l <loc>] [-t <tag>] # list items, optionally filtered
allourthings get <id-or-name> # show full item detail
allourthings add <name> [options] # add a new item
allourthings update <id> [options] # update item fields
allourthings delete <id> # delete an item (prompts for confirmation)
Attachment management:
allourthings attach add <item-id> <file> # attach a local file to an item
allourthings attach url <item-id> <url> # download a file and attach it
allourthings attach get <item-id> <filename> # save an attachment to disk
allourthings attach rm <item-id> <filename> # delete an attachment
add and update options:
-c, --category <category>
-b, --brand <brand>
-m, --model <model>
--purchase-date <date> ISO date, e.g. 2024-01-15
--price <price>
--currency <currency> e.g. GBP, USD
--warranty <date> warranty expiry ISO date
--retailer <retailer>
-l, --location <location>
--serial <serial>
-t, --tag <tag...> repeatable
-n, --notes <notes>
--set key=value custom/extra fields (update only, repeatable)
Global options:
--data-dir <path> path to inventory data directory (default: ~/Documents/AllOurThings)
--json output raw JSON — useful for scripting and agent use
Data directory: defaults to ~/Documents/AllOurThings on all platforms. To avoid passing --data-dir every time, set it once in your shell profile:
export ALLOURTHINGS_DATA_DIR=~/Dropbox/AllOurThings
The directory is created automatically on first write. Read commands (list, search, get) return empty results against a missing directory rather than erroring.
Examples
# Add an item
allourthings add "Bosch Washing Machine" --brand Bosch --model "WGG244A9GB" \
--category appliance --location kitchen \
--purchase-date 2024-01-15 --price 649 --currency GBP \
--warranty 2026-01-15 --retailer "John Lewis"
# Search and pipe to jq
allourthings search "warranty" --json | jq '[.[] | {name, warranty_expires}]'
# Attach a manual
allourthings attach add 6164c373 ~/Downloads/bosch-manual.pdf --label "User manual"
# Update a field
allourthings update 6164c373 --warranty 2027-01-15
# Use a custom data directory
allourthings --data-dir ~/Dropbox/AllOurThings list
Development
Prerequisites
Install dependencies
bun install
Tasks
| Task | Description |
|---|---|
task dev |
Seed vault + open MCP Inspector — fastest way to test |
task dev:mcp |
Start MCP server in watch mode (stdio) |
task test:run |
Run automated tests |
task seed |
Append test items to dev vault |
task seed:reset |
Clear dev vault and re-seed |
task inspect |
Open MCP Inspector (dev mode, no build required) |
task inspect:prod |
Build, then open MCP Inspector against compiled dist |
task build |
Compile MCP server to dist/ |
task build:cli |
Compile CLI to dist/ |
task cli -- <args> |
Run CLI from source against dev vault, e.g. task cli -- list |
task typecheck |
Run TypeScript type checking |
task clean |
Remove dist/ |
task clean:vault |
Delete local dev vault |
All tasks use ./dev-vault by default. Override with DATA_DIR=/your/path task <command>.
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.