Tessie MCP Server
Enables AI assistants to control and monitor Tesla vehicles via natural language through the Tessie API, supporting battery, climate, security, charging, and more.
README
Tessie MCP Server
Control and monitor your Tesla from any MCP-compatible AI assistant — including Claude, Cursor, and more.
An MCP (Model Context Protocol) server that bridges AI assistants to the Tessie API, giving you natural-language control of your Tesla vehicle. Ask Claude to check your battery, preheat the cabin, lock the car, or pull charging history — all in plain English.
✨ Features
- 🔋 Battery & Range — level, estimated range, voltage, temperature, health/degradation
- 📍 Location — GPS coordinates and street address
- 🌡️ Climate Control — start/stop, set temperature, seat heat/cool, defrost, Keep/Dog/Camp mode, Bioweapon Defense
- 🔒 Security — lock/unlock, Sentry Mode, Valet Mode, Guest Mode, Speed Limit Mode
- ⚡ Charging — start/stop, set charge limit %, set amps, charge port control
- 🚗 Drive Data — history with distance, energy & speed; GPS path per trip; drive tags
- 📊 Charge History — sessions with cost tracking
- 🛻 Trunks & Covers — frunk, rear trunk, Cybertruck tonneau
- 🪟 Windows & Sunroof — vent or close
- 💡 Lights & Horn — flash lights, honk
- 🏠 HomeLink — trigger garage door opener
- 🔑 Remote Start — keyless driving (2-minute window)
- 📱 Software Updates — schedule or cancel
- 😴 Auto-Wake — commands automatically wake a sleeping vehicle before executing
- 💨 Fun — Boombox fart sounds 🐄
⚡ Quick Start
1. Get a Tessie Account & API Token
Tessie is the service that provides secure API access to your Tesla.
👉 Sign up for Tessie: share.tessie.com/v4Gklbe1U0b
Once signed in, get your API token at: dash.tessie.com/settings/api
2. Install & Build
git clone https://github.com/akrowczyk/tessie-mcp-server.git
cd tessie-mcp-server
npm install
npm run build
🔧 Configuration
Environment Variables
Authentication and preferences are configured entirely via environment variables — no secrets ever touch source code.
| Variable | Required | Description |
|---|---|---|
TESSIE_API_TOKEN |
✅ Yes | Your Tessie API token |
TESSIE_DEFAULT_VIN |
Optional | Default VIN when not specified per-call |
TESSIE_DISTANCE_FORMAT |
Optional | mi or km — applies to all distance tools |
TESSIE_TEMP_FORMAT |
Optional | f or c — applies to all temperature tools |
TESSIE_PRESSURE_FORMAT |
Optional | psi, bar, or kpa — applies to tire pressure |
Per-call parameters always override the env var defaults.
Claude Desktop
Add this block to your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
{
"mcpServers": {
"tessie": {
"command": "node",
"args": ["/absolute/path/to/tessie-mcp-server/dist/index.js"],
"env": {
"TESSIE_API_TOKEN": "your-tessie-api-token",
"TESSIE_DEFAULT_VIN": "your-17-digit-vin",
"TESSIE_DISTANCE_FORMAT": "mi",
"TESSIE_TEMP_FORMAT": "f",
"TESSIE_PRESSURE_FORMAT": "psi"
}
}
}
}
Note: Replace
/absolute/path/to/tessie-mcp-serverwith the actual path where you cloned this repo.
After saving, restart Claude Desktop.
Claude Code (CLI)
claude mcp add tessie -- node /absolute/path/to/tessie-mcp-server/dist/index.js
export TESSIE_API_TOKEN="your-tessie-api-token"
export TESSIE_DEFAULT_VIN="your-17-digit-vin"
export TESSIE_DISTANCE_FORMAT="mi"
export TESSIE_TEMP_FORMAT="f"
export TESSIE_PRESSURE_FORMAT="psi"
Cursor
Add to .cursor/mcp.json (project) or ~/.cursor/mcp.json (global):
{
"mcpServers": {
"tessie": {
"command": "node",
"args": ["/absolute/path/to/tessie-mcp-server/dist/index.js"],
"env": {
"TESSIE_API_TOKEN": "your-tessie-api-token",
"TESSIE_DEFAULT_VIN": "your-17-digit-vin",
"TESSIE_DISTANCE_FORMAT": "mi",
"TESSIE_TEMP_FORMAT": "f",
"TESSIE_PRESSURE_FORMAT": "psi"
}
}
}
}
Windsurf
Open Settings → MCP Servers → Add:
{
"tessie": {
"command": "node",
"args": ["/absolute/path/to/tessie-mcp-server/dist/index.js"],
"env": {
"TESSIE_API_TOKEN": "your-tessie-api-token",
"TESSIE_DEFAULT_VIN": "your-17-digit-vin",
"TESSIE_DISTANCE_FORMAT": "mi",
"TESSIE_TEMP_FORMAT": "f",
"TESSIE_PRESSURE_FORMAT": "psi"
}
}
}
Continue.dev
In ~/.continue/config.json:
{
"experimental": {
"modelContextProtocolServers": [
{
"transport": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/tessie-mcp-server/dist/index.js"],
"env": {
"TESSIE_API_TOKEN": "your-tessie-api-token",
"TESSIE_DEFAULT_VIN": "your-17-digit-vin",
"TESSIE_DISTANCE_FORMAT": "mi",
"TESSIE_TEMP_FORMAT": "f",
"TESSIE_PRESSURE_FORMAT": "psi"
}
}
}
]
}
}
😴 Auto-Wake
All vehicle commands automatically wake a sleeping vehicle before executing — no need to call wake_vehicle manually first.
- If the vehicle is already awake, commands execute instantly with zero added latency
- If the vehicle is asleep, the server sends a wake command and polls every 3 seconds until the vehicle is awake (up to 90 seconds), then sends your command
- Read-only data tools (
get_battery,get_location, etc.) use cached data by default and do not wake the vehicle
📡 MCP Resources
In addition to tools, this server exposes vehicle data as MCP Resources — structured data that AI clients can read as contextual documents:
| URI | Description |
|---|---|
tessie://vehicles |
Fleet overview — all vehicles on the account |
tessie://{vin}/state |
Full live vehicle state (charge, climate, drive, config) |
tessie://{vin}/status |
Current awake/asleep status |
Resources are accessible in Claude Desktop and other MCP clients that support the resources protocol.
🛠️ Available Tools
Vehicle Data
| Tool | Description |
|---|---|
get_vehicles |
List all vehicles on the account |
get_vehicle_state |
Full state — drive, charge, climate, config |
get_vehicle_status |
Check if vehicle is awake/asleep |
get_full_status |
Combined battery + location + state in one call |
get_battery |
Battery level, range, voltage, temperature |
get_battery_health |
Battery degradation over time |
get_location |
GPS coordinates and street address |
get_weather |
Weather at the vehicle's current location |
get_tire_pressure |
All four tire pressures |
get_consumption_since_charge |
Energy usage since last charge |
get_firmware_alerts |
Vehicle firmware alerts |
Drives & Charges
| Tool | Description |
|---|---|
get_drives |
Drive history with distance, energy, speed |
get_drive_path |
GPS path for drives |
set_drive_tag |
Tag drives as business/personal |
get_charges |
Charging session history |
set_charge_cost |
Set cost for a charge session |
get_idles |
Idle periods (parked, not charging) |
Commands (all auto-wake)
| Tool | Description |
|---|---|
wake_vehicle |
Explicitly wake from sleep |
lock / unlock |
Door locks |
open_front_trunk / open_rear_trunk |
Trunk controls |
vent_windows / close_windows |
Window controls |
vent_sunroof / close_sunroof |
Sunroof controls |
open_tonneau / close_tonneau |
Cybertruck tonneau cover |
flash_lights / honk |
Lights and horn |
trigger_homelink |
Garage door opener |
remote_start |
Keyless driving (2-min window) |
boombox |
External speaker fart sound 🐄 |
Climate (all auto-wake)
| Tool | Description |
|---|---|
start_climate / stop_climate |
Climate system |
set_temperatures |
Set cabin temp (15–28°C) |
set_seat_heat / set_seat_cool |
Seat heating/cooling (0–3) |
start_defrost / stop_defrost |
Max defrost mode |
start_steering_wheel_heater / stop_steering_wheel_heater |
Steering wheel heater |
set_cabin_overheat_protection |
Cabin Overheat Protection |
set_climate_keeper_mode |
Keep / Dog / Camp mode |
set_bioweapon_mode |
Bioweapon Defense Mode |
Charging (all auto-wake)
| Tool | Description |
|---|---|
start_charging / stop_charging |
Start/stop charging |
set_charge_limit |
Set charge limit (50–100%) |
set_charging_amps |
Set charging amperage |
open_charge_port / close_charge_port |
Charge port door |
Modes (all auto-wake)
| Tool | Description |
|---|---|
enable_sentry_mode / disable_sentry_mode |
Sentry Mode |
enable_valet_mode / disable_valet_mode |
Valet Mode |
enable_guest_mode / disable_guest_mode |
Guest Mode |
enable_speed_limit / disable_speed_limit |
Speed Limit Mode |
Software (all auto-wake)
| Tool | Description |
|---|---|
schedule_software_update |
Schedule a software update |
cancel_software_update |
Cancel a scheduled update |
💬 Example Prompts
Once configured, try asking your AI assistant:
- "How is my Tesla doing right now?" → uses
get_full_status - "What's my battery level and estimated range?"
- "Where is my car parked right now?"
- "Start the climate and set it to 72°F" (auto-wakes if needed)
- "Lock the car and turn on Sentry Mode" (auto-wakes if needed)
- "Show me my drives from this week"
- "Open the frunk"
- "What's the tire pressure on all four tires?"
- "Set my charge limit to 80%"
- "Tag all my drives this month as business"
🔒 Security Notes
- Your Tessie API token is passed as an environment variable — never stored in code or config files
- The
.gitignorein this repo explicitly excludes.envfiles - If your token has been exposed, rotate it immediately at dash.tessie.com/settings/api
- All vehicle commands are authenticated with your personal token — only you can control your vehicle
🛠️ Development
npm install # Install dependencies
npm run build # Compile TypeScript
npm run dev # Watch mode (auto-rebuild on changes)
npm start # Run the compiled server
Requirements: Node.js 18+
🔗 Referrals & Links
If this project is useful to you, consider using these referral links:
- Tessie — The vehicle API that powers this server. Sign up here: share.tessie.com/v4Gklbe1U0b
- Buying a Tesla? Use my referral code for exclusive deals and rewards: ts.la/andrew80231
📄 License
MIT © 2026 Andrew Krowczyk
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。