convez.ai
MCP ServerAutomaçãoIntermediário
ober37

AC Infinity MCP – Controle de Ambientes de Cultivo

ober37/ac-infinity-mcp

Monitore e controle seu ambiente de cultivo por IA com dados ao vivo

Ver no GitHub
BomPluginScore · de 0 a 100

Servidor MCP que integra controladores AC Infinity (grow lights, ventiladores, umidificadores) com Claude e assistentes IA. Monitora temperatura, umidade e VPD em tempo real, controla portas via linguagem natural e automatiza o ambiente de cultivo sem abrir o app proprietário.

Instalação

$npx convez-ai ober37/ac-infinity-mcp

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
11
Forks
5
Acompanhando
0
Saúde
100%
Último commit
há 3 semanas
Versão
—
Issues abertas
64
Contribuidores
2

Casos de uso

  • Automação de ambientes de cultivo indoor com leitura de sensores em tempo real
  • Análise de tendências VPD e recomendações de ajuste automático
  • Controle remoto de ventiladores, umidificadores e ports via linguagem natural

Veredito

Ferramenta sólida e bem documentada para growers que usam AC Infinity. A integração MCP é limpa, suporta dry-run para segurança, e os tools cobrem leitura, análise e escrita. Ausência de suporte completo para 89 AI+ é uma limitação, mas v1 já entrega valor real.

README

A tradução para o português deste README ainda está na fila.


What is this?

ac-infinity-mcp is an MCP server that connects Claude (and other AI assistants) directly to your AC Infinity controllers — so you can read live sensor data, run analytics, and adjust fan speeds or port states using natural language, without opening the AC Infinity app.

Built with FastMCP and the AC Infinity cloud API.

Example prompts

"What's the VPD trend in my tent over the last 7 days?" "Is my environment in the right range for late flower?" "Turn off port 3 on my 69 Pro controller — dry run first." "Show me which ports have been running the most this week." "Apply the veg stage template to port 1 — dry run first so I can see the settings." "My VPD is too high — what should I adjust?"

Compatible hardware

ControllerReadsWritesNotes
UIS Controller 69 Pro✅✅Legacy protocol
UIS Controller 69 Pro+✅✅Legacy protocol
UIS Controller 89 AI+✅⚠️ v1 read-onlyNew AI+ protocol — write support planned for v2

Tools

CategoryToolWhat it does
🔍 Discoverydiscover_devicesList all your controllers and ports
🌡️ Readingsget_device_readingLive temp, humidity, VPD, and port states for one device
🌡️ Readingsget_all_device_readingsSame, for all devices at once
🌡️ Readingsget_historical_readingsTime-series data with optional bucketing (raw, 15m, 1h, 1d, …)
📊 Analyticscheck_vpd_driftVPD target compliance check for your current grow stage
📊 Analyticsget_environment_healthComposite health score (0–100, A–F) across temp, humidity, VPD
📊 Analyticsdetect_environment_trendsLinear trend + 7-day projection per metric
📊 Analyticsget_port_activity_reportPer-port on/off hours, uptime %, and peak activity hour
🔎 Port Statusget_port_statusLive port power level, load detection, active mode, and timer countdown
🔎 Port Statusget_port_settingsFull automation config: mode, VPD target, temp/humidity thresholds, schedule, cycle
🔌 Writeset_port_speedSet fan or pump speed 1–10 (dry_run=True by default)
🔌 Writeset_port_onTurn a port fully on (dry_run=True by default)
🔌 Writeset_port_offTurn a port fully off (dry_run=True by default)
⚙️ Automationset_vpd_automationEnable VPD mode with a kPa target using built-in sensors (dry_run=True by default)
⚙️ Automationset_temperature_automationEnable AUTO mode with min/max °C thresholds (dry_run=True by default)
⚙️ Automationset_humidity_automationEnable AUTO mode with min/max % RH thresholds (dry_run=True by default)
⚙️ Automationset_port_modeSwitch to any mode: OFF, ON, AUTO, VPD, CYCLE, SCHEDULE, TIMER_TO_ON, TIMER_TO_OFF (dry_run=True by default)
🌱 Intelligenceapply_grow_stage_templateOne-click VPD + temp + humidity automation for a named grow stage (dry_run=True by default)
🤖 Advance Automationlist_advance_automationsList all named Advance Automation programs on a device
🤖 Advance Automationget_advance_automationGet full detail (schedule, port groups, run state) for one automation
🤖 Advance Automationenable_advance_automationEnable a disabled automation — reads state first, no-ops if already enabled (dry_run=True by default)
🤖 Advance Automationdisable_advance_automationDisable an enabled automation — reads state first, no-ops if already disabled (dry_run=True by default)
🤖 Advance Automationcreate_advance_automationCreate a new named automation with speed, schedule, and threshold settings (dry_run=True by default)
🤖 Advance Automationdelete_advance_automationDelete an automation (disables first if active) (dry_run=True by default)
🤖 Advance Automationbreak_out_of_automationSafely break a port out of automation control and lock co-governed ports to manual speed (dry_run=True by default)

✦ All write tools (Write, Automation, and Intelligence categories) default to dry_run=True — they return the exact payload they would send without making any changes to your equipment. Pass dry_run=False only when you're ready to execute.

📖 For a complete grower's guide with conversation examples for every tool, see docs/GUIDE.md.

MCP Prompts

Three built-in prompts are registered alongside the tools. In Claude Desktop and other MCP clients, these appear as slash commands or prompt suggestions.

PromptWhat it does
vpd_troubleshootingStep-by-step guide: diagnose HIGH or LOW VPD and which tools to call to fix it
new_grower_setupOnboarding walkthrough: discover devices → apply a stage template → check your health score
environment_alert_interpretationHow to read check_vpd_drift status values and get_environment_health grades (A–F, score weighting, what to do)

Quick start

Option 1: pip (simplest)

pip install git+https://github.com/ober37/ac-infinity-mcp.git

Requires Python 3.11+.

Option 2: isolated venv (recommended)

python3 -m venv ~/.venvs/ac-infinity-mcp
source ~/.venvs/ac-infinity-mcp/bin/activate
pip install git+https://github.com/ober37/ac-infinity-mcp.git
which ac-infinity-mcp   # copy this path — you'll need it below

Option 3: uvx (no install)

uvx --from git+https://github.com/ober37/ac-infinity-mcp.git ac-infinity-mcp

MCP client configuration

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "ac-infinity": {
      "command": "/path/to/ac-infinity-mcp",
      "env": {
        "AC_INFINITY_EMAIL": "you@example.com",
        "AC_INFINITY_PASSWORD": "yourpassword"
      }
    }
  }
}

Replace /path/to/ac-infinity-mcp with the path printed by which ac-infinity-mcp.

Cursor / Windsurf

Add to your MCP settings:

{
  "ac-infinity": {
    "command": "/path/to/ac-infinity-mcp",
    "env": {
      "AC_INFINITY_EMAIL": "you@example.com",
      "AC_INFINITY_PASSWORD": "yourpassword"
    }
  }
}

Docker

cp .env.example .env
# fill in AC_INFINITY_EMAIL and AC_INFINITY_PASSWORD
docker compose up --build

The container runs over stdio. Connect via a stdio bridge such as mcp-proxy.

Credentials are injected at runtime via env_file — never baked into the image.

Environment variables

VariableRequiredDescription
AC_INFINITY_EMAILYesYour AC Infinity account email
AC_INFINITY_PASSWORDYesYour AC Infinity account password

⚠️ HTTP-only upstream: The AC Infinity cloud API does not support HTTPS. Credentials and sensor data traverse the network in plain text. This is an upstream limitation — see docs/API.md Quirk 8. Avoid running this server on untrusted networks.

Development

git clone https://github.com/ober37/ac-infinity-mcp.git
cd ac-infinity-mcp
python3 -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
cp .env.example .env   # fill in credentials

# Lint + type-check
ruff check src/ tests/
mypy src/ac_infinity_mcp/

# Tests (unit + integration)
pytest tests/ -v

# Security audit
pip-audit

Security

Vulnerabilities should be reported via GitHub Private Vulnerability Reporting. See SECURITY.md for the full disclosure policy, scope, and known limitations (including the HTTP-only upstream API). Accepted dependency CVEs are tracked in docs/SECURITY-RISKS.md.

License

MIT

Ver o original no GitHub

Bom para

Relacionados