nest-logger

Log historical temperature readings from all Nest room sensors and graph them over time. Each sensor appears as its own line on a time-series chart, with thermostat HVAC state (heating / cooling / idle) shown as background shading. Annotate your data directly in the chart.

How it works

A Chrome extension scrapes the home.nest.com web app every 5 minutes and POSTs the readings to a small Express server running on your local machine. The server stores everything in an SQLite database and serves an interactive Plotly.js chart.

home.nest.com tab
  └─ Chrome extension (scraper.js + background.js)
       └─ POST /api/readings
            └─ Express server (server.js)
                 ├─ SQLite database (nest.db)
                 └─ Plotly.js chart (localhost:51920)

Requirements

• Node.js 18+ (or use nvm)
• Google Chrome
• A Nest account with at least one thermostat, accessible at home.nest.com
• Linux with systemd (for the optional always-on service)

Setup

1. Clone the repo

git clone https://github.com/joeyparrish/nest-logger.git
cd nest-logger

2. Start the server

cd server
npm install
npm start

The server listens on port 51920. Open http://localhost:51920 to see the chart (it will be empty until the extension sends its first reading).

3. Load the Chrome extension

1. Open chrome://extensions in Chrome.
2. Enable Developer mode (toggle in the top-right corner).
3. Click Load unpacked and select the extension/ folder from this repo.

Once loaded, navigate to home.nest.com in a tab and log in. The extension will automatically navigate to your thermostat page and begin scraping every 5 minutes. You should see readings appear in the chart within a few minutes.

Note: The tab must remain open for data to be collected. The server runs independently and does not need the tab to serve the chart. The extension monitors itself: if no reading arrives for more than 10 minutes it automatically reloads the tab, and the tab is also reloaded once a day around 3 am to clear accumulated memory, just in case of an unknown leak.

Running the server at boot (Linux + systemd)

To keep the server running in the background without leaving a terminal open, install it as a systemd user service. This also starts it automatically at boot without requiring a login.

cp server/nest-logger.service ~/.config/systemd/user/

Edit ~/.config/systemd/user/nest-logger.service and update the ExecStart line to the absolute path of start.sh in your clone:

ExecStart=/bin/bash /path/to/nest-logger/server/start.sh

Then enable and start the service:

loginctl enable-linger $USER   # allow user services to run without login (once)
systemctl --user daemon-reload
systemctl --user enable --now nest-logger

Useful commands:

systemctl --user status nest-logger
journalctl --user -u nest-logger -f   # live logs
systemctl --user restart nest-logger
systemctl --user stop nest-logger

nvm users: start.sh automatically loads nvm and uses whatever Node version is set as your default. No need to hard-code a version.

Chart features

• Temperature lines — one per sensor, rendered with WebGL for smooth panning and zooming over large date ranges.
• HVAC shading — semi-transparent red/blue background bands show when the system was heating or cooling.
• Range selector buttons — quickly jump to the last 12 hours, 1 day, 1 week, 1 month, or the full history.
• Range slider — drag the handles below the timeline to pan and zoom.
• Status bar — shows the number of readings in view and time spent heating, cooling, and idle for the visible range.
• Last reading indicator — shows how long ago the most recent data point was collected, updated every minute.
• Auto-refresh — the chart fetches new data automatically, timed to arrive shortly after each new reading is expected from the extension.
• Annotations — Ctrl+click (or ⌘+click on Mac) on any data point opens an annotation form pre-filled with the clicked timestamp. Annotations appear as dotted vertical lines with triangle markers; hovering shows the note text.

Development

To populate the database with synthetic data for testing without the extension:

cd server
node seed.js

This generates 3 days of 5-minute readings with a sinusoidal temperature curve and simulated HVAC cycles, inserted with INSERT OR IGNORE so it is safe to run against a database that already contains real data (seed timestamps are set to one year before the tool was created).

Data

The SQLite database is stored at server/nest.db and is created automatically on first run. It is not checked into the repository. Back it up if you want to preserve your history.

sensor_readings (timestamp, section, sensor, value)
hvac_states     (timestamp, action)

All timestamps are UTC ISO 8601 strings. The chart converts them to the browser's local timezone for display.