Installation - KnevS/Tesla-Carview GitHub Wiki

Installation Guide

🌐 Language: EN · DE · FR · ES · TR · EL

Time needed: ~30 minutes | Difficulty: Beginner-friendly

This guide walks you through a complete Tesla Carview installation from scratch.

What you need before you start

Before touching any commands, make sure you have:

  • A Linux server, VPS, or Raspberry Pi (see hardware options below)
  • A domain name pointing to your server — OR plan to use DynDNS / Cloudflare Tunnel (→ Network Access)
  • A Tesla Developer account (→ Tesla API Setup)
  • SSH access to your server (or a keyboard + screen on the Pi)

Hardware options

Option 1: Raspberry Pi (home server)

Best for: personal use at home, low cost (~€60–120 total)

Model RAM Recommended storage
Raspberry Pi 5 (recommended) 4 GB or 8 GB NVMe SSD via M.2 HAT+
Raspberry Pi 4 4 GB USB SSD
Raspberry Pi 3 1 GB no longer supported ¹

¹ Raspberry Pi 3 and older (32-bit ARM) are no longer supported as of v3.51.0: Node.js ships no ARMv7 images from version 24 onwards.

⚠️ Important: Do not use an SD card for permanent operation. It will fail within months under Tesla Carview's write load. See Raspberry Pi Storage for a 20-minute fix.

Option 2: VPS at a hosting provider

Best for: 24/7 availability, no hardware to manage, easy setup

Provider Monthly cost Notes
Hetzner CX22 ~€4.35 Recommended, very reliable
netcup VPS 1000 ~€4.44 German data centres
Contabo VPS S ~€5.99 Lots of storage

A VPS has a fixed public IP address — no DynDNS setup needed.

Step 1: Prepare the server

Connect to your server via SSH (or open a terminal on the Pi):

ssh root@YOUR-SERVER-IP

Make sure the system is up to date:

apt update && apt upgrade -y

Step 2: Get a domain pointing to your server

Tesla Carview requires HTTPS (Tesla's API only works over secure connections). That means you need a domain with a valid SSL certificate.

I have a VPS with a fixed IP: → Go to your domain registrar and create an A record:

tesla.yourdomain.com  →  A  →  YOUR-VPS-IP

Wait 5–30 minutes for DNS to propagate, then continue.

I'm at home without a fixed IP: → See Network Access — set up DynDNS or Cloudflare Tunnel first, then come back here.

I have no domain at all: → Get a free subdomain at DuckDNS.org (e.g. my-tesla.duckdns.org) — it's free and works with Let's Encrypt.

Step 3: Run the setup script

This single command downloads and runs the interactive setup wizard:

curl -fsSL https://raw.githubusercontent.com/KnevS/Tesla-Carview/main/deploy/setup.sh | bash

The wizard asks you a series of questions:

Question What to enter
Domain name tesla.yourdomain.com or my-tesla.duckdns.org
Admin username Any name (e.g. your first name, admin)
Admin password A strong password (min. 12 characters)
Tenant name What to call your installation (e.g. "My Tesla")
Enable HTTPS Yes (always — required for Tesla API)

The script asks for the deployment mode before doing anything:

Mode When to use
1 — Direct Dedicated VM or VPS without an existing proxy — nginx + certbot are installed and configured automatically
2 — Proxy Already running nginx, Caddy, Traefik, or WireGuard+VPS setup — nginx installation is skipped entirely

In Proxy mode, add this to your existing proxy config:

proxy_pass http://127.0.0.1:8080;

⚠️ In proxy mode you have to reproduce the rate limits yourself — setup.sh creates no nginx configuration in this mode. If they are missing or too tight, the app runs into HTTP 429: half-loaded pages that only a reload appears to fix.

Path Recommendation Why
/api/auth/login 10/min, burst 3 Brute-force protection
/api/tiles/ 1200/min, burst 300 One map zoom loads 50–150 tiles at once
/api/ (rest) 120/min, burst ≥ 60 One page change fires 15–26 requests

The more specific path has to win — otherwise a single map zoom locks out the entire API. Templates: deploy/nginx-host.conf.template and deploy/traefik-dynamic.example.yml.

The script then automatically (Direct mode):

  1. Installs Docker, nginx, certbot, fail2ban
  2. Gets a Let's Encrypt SSL certificate for your domain
  3. Configures nginx with security headers
  4. Starts all Docker containers
  5. Sets up the database

This takes 5–10 minutes.

Step 4: First login

Open your browser and go to https://tesla.yourdomain.com

You should see the Tesla Carview login page. Enter the admin username and password you set in Step 3.

💡 Tip: Check "Stay signed in (90 days)" on the login page so you don't have to type your password every time — especially useful when accessing from the Tesla browser.

Step 5: Connect your Tesla account

After logging in, you'll see a setup prompt to connect your Tesla account. Follow the instructions in Tesla API Setup.

Step 6: Done!

Your Tesla Carview installation is running. The app will start polling your vehicle's data automatically.

What to do next:

  • Set up your vehicle → Dashboard → Add Vehicle
  • Configure notifications → Settings → Push Notifications
  • Invite family members → Admin → Users → Invite
  • Set up a charging location → Settings → Charging Locations

Updating

Tesla Carview can update itself automatically. Enable it in the settings:

# In /opt/tesla-carview/backend/.env:
AUTO_UPDATE_ENABLED=true

Or update manually:

cd /opt/tesla-carview
git pull
docker compose -f docker-compose.prod.yml up -d --build

Troubleshooting installation issues

"Permission denied" when running the script → Make sure you're running as root. Run sudo su first.

"Domain not found" during certbot → Your DNS hasn't propagated yet. Wait 10–30 minutes and try again. Check with: nslookup tesla.yourdomain.com

Containers won't start → Check logs: docker compose -f /opt/tesla-carview/docker-compose.prod.yml logs

More help → Troubleshooting | Open an Issue

Windows PC instead of a server (from v3.53.0)

The application runs entirely in Linux containers — Windows is only the host. With Docker Desktop in WSL2 mode the same stack starts. The setup script deploy/setup.sh does not work there (bash, apt, systemd, certbot); deploy\setup-windows.ps1 takes its place.

git clone https://github.com/KnevS/Tesla-Carview.git
cd Tesla-Carview
powershell -ExecutionPolicy Bypass -File .\deploy\setup-windows.ps1

Two limitations, both measured: vehicle commands do not work (tesla-http-proxy needs a Linux bind mount and a fixed UID), and if you run the proxy anyway you must set TESLA_PROXY_BASE — under Docker Desktop host.docker.internal points at the host gateway instead of the proxy container. Everything read-only works fully.

Windows is not checked in CI. Full description: docs/02-deployment.en.md, section "Windows (Docker Desktop)".