MiniMed Connect is a device which communicates with a Medtronic insulin pump over 916MHz radio to gather data about pump status. This data may include readings from an Enlite continuous glucose monitor (CGM). The Connect forwards pump data via Bluetooth LE to a proprietary app running on an iPhone. The iPhone app uploads the data to Medtronic's servers, enabling authorized users to view it on the Medtronic CareLink Connect website.
Nightscout is an open source, DIY project that allows real-time access to CGM data on any platform. The core of the Nightscout project is cgm-remote-monitor, a web service which can receive and broadcast CGM data using a Mongo database. Various other software projects communicate with an instance of cgm-remote-monitor
, either to transmit data from a CGM device, or to display that CGM data on a computer, smartphone/tablet, or smartwatch.
minimed-connect-to-nightscout is a web scraper which sends data from Medtronic CareLink Connect to Nightscout. It does so by posing as a web browser, logging into CareLink Connect, periodically downloading pump status data from the Medtronic server, and uploading that data to a Nightscout server.
- A MiniMed Connect and compatible Medtronic insulin pump
- An iPhone running the MiniMed Connect app
- Username and password for a CareLink account linked to the Connect
- A working Nightscout website and Mongo database
The easiest installation mode is to set up an instance of Nightscout cgm-remote-monitor on Azure and enable the mmconnect
plugin. This module is packaged with Nightscout 0.8.2+ and can pull data from CareLink Connect as part of the web server process. Follow this guide.
Another turnkey installation option is to run this on a Heroku worker dyno. You may find this more reliable than Azure. Follow the Share2 Bridge instructions for Heroku, substituting this repo for share2nightscout-bridge
.
- Install Node.
- Clone this repository or download a zip with the latest version.
npm install
to install dependencies.- Set environment variables (see below).
npm start
and leave it running.
CARELINK_USERNAME
- your username for CareLinkCARELINK_PASSWORD
- your password for CareLinkAPI_SECRET
- the value you use forAPI_SECRET
on your Nightscout websiteWEBSITE_HOSTNAME
- the hostname for your Nightscout instance, which looks likeyour.host.com
. If you are running this script in the same Azure environment as Nightscout, there is no need to set this, as it will already be set by Azure. If you setNS
(see below), you do not need to set this.
CARELINK_REQUEST_INTERVAL
- number of milliseconds to wait between requests to the CareLink server (default: 60000)CARELINK_SGV_LIMIT
- maximum number of recent sensor glucose values to send to Nightscout (default: 24)CARELINK_MAX_RETRY_DURATION
- maximum number of seconds to spend retrying failed requests to CareLink, ideally a power of 2 (default: 512)CARELINK_QUIET
- set to a truthy value to not output details of CareLink and Nightscout requests to the console (default: empty)NS
- a fully-qualified Nightscout URL (e.g.https://sitename.azurewebsites.net
) which overridesWEBSITE_HOSTNAME
- Sensor glucose values and trend (single/double arrow up/down)
- Pump: active insulin, reservoir level, battery level
- MiniMed Connect: battery level, connection status to phone, connection status to pump
- Sensor: calibration state, time until next calibration, sensor duration, connection status to pump
- Pump model
Understanding of the current data is based mostly on this analysis.
run.js demonstrates how to use the key API features in production. A minimal example would look like:
var mmcns = require('minimed-connect-to-nightscout');
var client = mmcns.carelink.Client({username: 'username', password: 'password'});
client.fetch(function(err, data) {
if (!err) {
var transformed = mmcns.transform(data);
mmcns.nightscout.upload(transformed.entries, 'https://your.ns.host/api/v1/entries.json', 'api-secret', callback);
// ...or:
mmcns.nightscout.upload(transformed.devicestatus, 'https://your.ns.host/api/v1/devicestatus.json', 'api-secret', callback);
// ...or use `transformed.entries` and `transformed.devicestatus` directly
}
});
File an issue if you'd like to give feedback, request an enhancement, or report a bug.
Pull requests are welcome, provided they include tests. See test/
for examples. Run npm test
to run the suite.
Much of the Medtronic HTTP interaction is based on the excellent work by @bewest and @ianjorgensen on mmcsv.
This project is intended for educational and informational purposes only. It relies on a series of fragile components and assumptions, any of which may break at any time. It is not FDA approved and should not be used to make medical decisions. It is neither affiliated with nor endorsed by Medtronic, and may violate their Terms of Service.