Skip to main content
Glama
mouldiwarp

solax-cloud-mcp

by mouldiwarp

SolaX Cloud MCP Server

An MCP server that provides real-time access to solar inverter data from the SolaX Developer Platform API. Query your inverter's current power output, energy yields, battery status, and grid import/export data directly from Claude Code or Claude Desktop.

Prerequisites

  • Python 3.10+ (automatically provisioned by uv)

  • uv package manager (install here)

  • A SolaX Developer Platform account with OAuth2 application registered

  • Device (inverter) serial number (deviceSn) for your inverter

Related MCP server: EcoFlow MCP Server

Getting Started

1. Register an OAuth2 Application

  1. Log in to SolaX Developer Platform

  2. Navigate to Application section

  3. Create a new application and enable client_credentials grant type

  4. Copy your Client ID and Client Secret (keep these secret!)

2. Identify Your Device Serial Number

  1. Log in to SolaX Developer Platform

  2. Navigate to My Account or device management section

  3. Find your inverter's device serial number (e.g., X3ABCD0123)

    • This is NOT the old WiFi dongle registration number used by the legacy SolaX Cloud API

3. Install the Server

# Clone or navigate to the repo
cd /Users/gary/mysrc/claude/solax-cloud-mcp

# Install uv if needed
curl -LsSf https://astral.sh/uv/install.sh | sh

# Install dependencies and create virtual environment
uv sync

4. Configure Credentials

# Copy the example environment file
cp .env.example .env

# Edit .env and fill in your credentials
nano .env

Add your OAuth2 credentials and device serial number:

SOLAX_CLIENT_ID=your_client_id
SOLAX_CLIENT_SECRET=your_client_secret
SOLAX_DEVICE_SN=X3ABCD0123

5. Register with Claude Code

To use this server with Claude Code or Claude Desktop:

claude mcp add solax-cloud \
  --env SOLAX_CLIENT_ID=your_client_id \
  --env SOLAX_CLIENT_SECRET=your_client_secret \
  --env SOLAX_DEVICE_SN=your_device_sn \
  --scope user \
  -- uv run --directory /Users/gary/mysrc/claude/solax-cloud-mcp solax-cloud-mcp

If the above doesn't work (environment variable propagation issues), edit your MCP config JSON directly:

  • Claude Code: ~/.claude/mcp.json or project settings

  • Claude Desktop: ~/.config/Claude/claude_desktop_config.json (macOS) or equivalent

Add this entry:

{
  "solax-cloud": {
    "command": "uv",
    "args": [
      "run",
      "--directory",
      "/Users/gary/mysrc/claude/solax-cloud-mcp",
      "solax-cloud-mcp"
    ],
    "env": {
      "SOLAX_CLIENT_ID": "your_client_id",
      "SOLAX_CLIENT_SECRET": "your_client_secret",
      "SOLAX_DEVICE_SN": "your_device_sn"
    }
  }
}

Usage

Once registered, the get_realtime_data tool is available in your MCP toolset. Use it to query real-time inverter data.

Tool: get_realtime_data

Arguments:

  • device_sn (optional): Inverter device serial number. If omitted, defaults to SOLAX_DEVICE_SN environment variable.

Returns: A structured dictionary containing:

{
  "device": {
    "deviceSn": "X3ABCD0123",
    "registerNo": "SE123456SP",
    "dataTime": "2025-05-22 15:13:10",
    "plantLocalTime": "2025-05-22 15:13:10"
  },
  "status": {
    "code": 102,
    "description": "Normal"
  },
  "pv": [
    {"string": 1, "voltage_V": 421.7, "current_A": 0.4, "power_W": 189.0},
    {"string": 2, "voltage_V": 418.0, "current_A": 0.0, "power_W": 0.0}
  ],
  "mppt": {
    "trackers": [
      {"mppt": 1, "voltage_V": 0.0, "current_A": 0.0, "power_W": 0.0}
    ],
    "totalPower_W": null
  },
  "ac": {
    "phases": [
      {"phase": 1, "voltage_V": 224.1, "current_A": 0.9, "power_W": 189.0, "frequency_Hz": 50.0},
      {"phase": 2, "voltage_V": 226.5, "current_A": 0.8, "power_W": 171.0, "frequency_Hz": 50.0}
    ],
    "totalPower_W": 360.0,
    "totalReactivePower": 0,
    "powerFactor": 1.0,
    "gridFrequency": 50.0
  },
  "energy": {
    "dailyYield_kWh": 157.4,
    "totalYield_kWh": 20465.3,
    "dailyACOutput_kWh": 160.9,
    "totalACOutput_kWh": 19907.5
  },
  "meter1": {
    "gridPower_W": 0,
    "todayImportEnergy_kWh": 0.07,
    "totalImportEnergy_kWh": 23.28,
    "todayExportEnergy_kWh": 0.0,
    "totalExportEnergy_kWh": 64.98
  },
  "meter2": {
    "gridPower_W": 0,
    "todayImportEnergy_kWh": 0.08,
    "totalImportEnergy_kWh": 1.39,
    "todayExportEnergy_kWh": 0.0,
    "totalExportEnergy_kWh": 0.75
  },
  "battery": {
    "soc_percent": 85.5,
    "remainingEnergy_kWh": 1024.8,
    "soh_percent": 99.2,
    "chargeDischargePower_W": -150.5,
    "voltage_V": 409.6,
    "current_A": -15.2,
    "temperature_C": 28.3,
    "cycleTimes": 142,
    "totalCharge_kWh": 4250.75,
    "totalDischarge_kWh": 4100.25,
    "status": {"code": 1, "description": "Work"}
  },
  "eps": {
    "voltage_V": [0.0, 0.0, 0.0],
    "current_A": [0.0, 0.0, 0.0],
    "activePower_W": [0.0, 0.0, 0.0],
    "apparentPower_W": [0.0, 0.0, 0.0]
  },
  "temperature": {
    "inverter_C": 40.4
  },
  "misc": {
    "l1l2Voltage_V": null,
    "l2l3Voltage_V": null,
    "l1l3Voltage_V": null
  }
}

Example Usage in Claude

"What's the current power output of my solar inverter?"

Claude will call get_realtime_data() and report the results to you in human-friendly terms.

Docker: Build, Deploy & HTTP Consumption

For containerized deployment on Raspberry Pi or any Docker-enabled system, you can run the server in HTTP mode:

Build the Docker Image

# Build the image
docker build -t solax-cloud-mcp:latest .

# Verify the build
docker images | grep solax-cloud-mcp

Deploy with Docker Compose

Configure your environment variables first:

# Copy and edit the environment file
cp .env.example .env
nano .env

Fill in your SolaX credentials and generate a strong API key:

SOLAX_CLIENT_ID=your_client_id
SOLAX_CLIENT_SECRET=your_client_secret
SOLAX_DEVICE_SN=your_device_sn
HTTP_API_KEY=$(python -c "import secrets; print(secrets.token_urlsafe(32))")

Start the HTTP server:

# Build and start the container in the background
docker-compose up -d

# Check logs
docker-compose logs -f solax-http

# Verify it's running
curl http://localhost:8000/health

The server listens on port 8000 and is accessible at http://YOUR_PI_IP:8000.

Consume the HTTP API

All endpoints (except /health) require a bearer token in the Authorization header.

Health Check (no authentication)

curl http://192.168.1.100:8000/health

Response:

{"status": "ok"}

Get Real-Time Inverter Data

curl -X POST \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"device_sn": "X3ABCD0123"}' \
  http://192.168.1.100:8000/api/realtime-data

Returns current power output, battery SOC, grid export/import, and more.

Set Battery Self-Use Mode

Configure battery charging/discharging thresholds:

curl -X POST \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "device_sn": "X3ABCD0123",
    "min_soc": 20,
    "charge_upper_soc": 80,
    "charge_from_grid_enable": 1
  }' \
  http://192.168.1.100:8000/api/battery/self-use-mode

Python / JavaScript Examples

Quick examples in Python and JavaScript are available in HTTP_API.md, including:

  • Polling for real-time updates

  • Time-based battery scheduling

  • Error handling patterns

  • Rate limiting considerations

Deployment on Raspberry Pi

Full deployment instructions (Docker installation, network setup, security, monitoring) are in DEPLOYMENT.md.

Docker Management

# View logs
docker-compose logs -f solax-http

# Restart the server
docker-compose restart solax-http

# Stop the server
docker-compose down

# Rebuild after code changes
docker-compose up -d --build

# Check resource usage
docker stats solax-http

Manual Testing

Quick Smoke Test

Before registering with Claude, you can test the client directly:

uv run --env-file .env python -c \
  "import asyncio; from solax_cloud_mcp.client import fetch_realtime_data; \
   from solax_cloud_mcp.config import get_default_device_sn; \
   print(asyncio.run(fetch_realtime_data(get_default_device_sn())))"

Run Tests

uv run pytest -v

Tests include:

  • Response shaping and status decoding

  • Environment variable validation

  • Error handling and edge cases (battery-less devices, null status codes)

  • Dynamic PV/MPPT parsing

  • Case-insensitive field access

Interactive Inspection

If you have the MCP CLI tools installed:

uv run --env-file .env mcp dev src/solax_cloud_mcp/server.py

Rate Limiting

This server respects SolaX Developer Platform's documented rate limits:

  • 100 calls per minute per token

  • 1,000,000 calls per day per token

The client automatically enforces a 0.7-second minimum spacing between calls, which keeps typical usage well within both limits. No manual rate limiting is needed.

Data Fields Reference

All values are in SI units:

  • Power: Watts (W)

  • Energy: Kilowatt-hours (kWh)

  • Voltage: Volts (V)

  • Current: Amps (A)

  • Temperature: Celsius (°C)

  • Frequency: Hertz (Hz)

Inverter Status Codes

The inverter's status.code is an integer from the table below (Appendix 6 from the SolaX Developer Platform API docs). Not all possible codes are listed; consult the full Appendix 6 table on developer.solaxcloud.com/doc for the complete set.

Code

Status

Description

100

Waiting

Waiting

101

Self-check

Self-check

102

Normal

Normal

103

Fault

Fault

104

Permanent Fault Mode

Permanent Fault Mode

105

Update Mode

Update Mode

106

EPS Check Mode

EPS Check Mode

107

EPS Mode

EPS Mode

108

Self-test

Self-test

109

Idle Mode

Idle Mode

110

Standby Mode

Standby Mode

130

VPP mode

VPP mode

131

TOU-Self use

TOU-Self use

132

TOU-Charging

TOU-Charging

133

TOU-Discharging

TOU-Discharging

1301-1309

Advanced control modes

Power/SOC target control, self-consume modes, etc.

Battery Status Codes (Residential)

The battery's battery.status.code is an integer:

Code

Status

0

Idle

1

Work

API Error Codes (Appendix 1)

Code

Message

10000

Operation successful

10001

Operation failed

11500

System busy, please try again later

10200

Operation abnormality, please see the specific message content for details

10400

Request not authenticated

10401

Username or password incorrect

10402

Request access_token authentication failed

10403

Interface has no access rights

10404

Callback function not configured

10405

The number of API calls has been used up

10406

The API call rate has reached the upper limit, please try again later

10500

User has no device data permission

10505

Device unauthorized

10506

Plant unauthorized

Troubleshooting

"SOLAX_CLIENT_ID environment variable not set"

  • Set the SOLAX_CLIENT_ID environment variable or register the server with the correct credentials

"SOLAX_CLIENT_SECRET environment variable not set"

  • Set the SOLAX_CLIENT_SECRET environment variable or register the server with the correct credentials

"No device_sn provided"

  • Either pass the device_sn argument to the tool or set the SOLAX_DEVICE_SN environment variable

"SolaX API error 10402: Request access_token authentication failed"

  • The server will automatically attempt to refresh the access token once; if this persists, your Client Secret may be invalid or revoked. Re-register your OAuth2 application at https://developer.solaxcloud.com/

"SolaX API error 10505: Device unauthorized"

  • The device serial number you provided is invalid or not associated with your account. Double-check the device SN in your SolaX Developer Platform account.

"SolaX API error 10406: The API call rate has reached the upper limit"

  • The rate limiter is correctly enforced; this should rarely occur under normal usage. If it does, the server automatically backs off. Reduce tool call frequency or wait a few seconds and retry.

"Network timeout"

  • SolaX Developer Platform API is unreachable. Check your internet connection and confirm the service is online at https://developer.solaxcloud.com/.

Development

Project Structure

solax-cloud-mcp/
├── src/solax_cloud_mcp/
│   ├── __init__.py         # Package metadata
│   ├── __main__.py         # CLI entry point
│   ├── server.py           # MCP server and tool definitions
│   ├── client.py           # HTTP client and API calls
│   ├── auth.py             # OAuth2 token management
│   ├── config.py           # Environment variable handling
│   └── models.py           # Data models and response shaping
├── tests/
│   ├── fixtures/           # Test data
│   ├── test_config.py      # Configuration tests
│   └── test_models.py      # Response shaping tests
├── pyproject.toml          # Project metadata and dependencies
└── README.md               # This file

Adding Features

The server is designed to be minimal and focused. To add more endpoints/tools:

  1. Fetch the data from SolaX Developer Platform API (extend client.py)

  2. Add a response shaping function if needed (extend models.py)

  3. Define a new @server.tool() in server.py

License

MIT

Support

For issues with this MCP server, open an issue on the repository.

For SolaX Developer Platform API documentation, refer to SolaX Developer Platform.

Available Tools

2 tools
get_realtime_dataA

Fetch real-time solar inverter data from SolaX Developer Platform.

Retrieves the latest readings including power output, energy yields, battery status, and grid import/export data from the SolaX Developer Platform API. All power values are in Watts (W), energy in kWh, voltage in Volts (V), current in Amps (A), and temperature in Celsius (°C).

Args: device_sn: Serial number of the inverter to query (the device SN, not WiFi dongle SN). If omitted, defaults to SOLAX_DEVICE_SN environment variable.

Returns: A structured dictionary with decoded status values and grouped data fields: - device: inverter identifiers (SN, registration number, timestamps) - status: inverter operating status (code + human-readable description) - pv: list of active PV strings with voltage/current/power - mppt: MPPT tracker data (individual trackers + total power) - ac: AC output phases, total power, reactive power, power factor, grid frequency - energy: daily/total yield and AC output - meter1: grid power and energy (import/export) - meter2: secondary meter data (import/export) - battery: battery voltage/current/power/soc/status (None if no battery) - eps: emergency power supply data (3-phase voltage/current/power) - temperature: inverter temperature - misc: line-to-line voltages

Raises: ToolError: if credentials are invalid, the device is not found, or the API fails.

ParametersJSON Schema
NameRequiredDescriptionDefault
device_snNo

TDQS

A4.6/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Describes the operation as read-only ('Fetch', 'Retrieves'), lists error conditions, and provides return structure details. No annotations are present, so description covers the behavioral aspects well.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Well-structured with summary, units, clear Args/Returns/Raises sections. Front-loaded with main purpose, no superfluous content.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a single-parameter tool, the description fully covers input, output (detailed return structure), and error behavior, making it self-contained without needing an output schema.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Clearly explains the parameter 'device_sn', including the distinction from WiFi dongle SN and the default environment variable. This adds significant value beyond the schema which lacks descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states 'Fetch real-time solar inverter data' with a specific data source. It distinguishes from the sibling tool 'set_battery_self_use_mode' which is a write operation.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Provides context like data source and default environment variable, but lacks explicit when-to-use or when-not-to-use guidance relative to siblings.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

set_battery_self_use_modeA

Set inverter battery to Self Use Mode with configurable charging thresholds.

Self Use Mode is ideal for automated control based on conditions like weather. You can set minimum SOC (don't discharge below), maximum charge SOC (don't charge above), and optionally configure time-based charge/discharge periods (e.g., charge during peak solar, discharge during high-rate periods).

Example automation scenarios:

  • Reduce charging threshold on cloudy days to preserve grid import reserves

  • Increase charging threshold on clear days to maximize solar self-consumption

  • Prevent overnight charging from grid during expensive peak hours

Args: device_sn: Serial number of the inverter. If omitted, uses SOLAX_DEVICE_SN. min_soc: Minimum SOC (%), range [10, 100]. Battery won't discharge below this. Default: 10%. charge_upper_soc: Maximum charging SOC (%), range [10, 100]. Battery won't charge above this. Default: 100%. charge_from_grid_enable: Allow charging from grid (0=no, 1=yes). Default: 1 (enabled). charge_start_time_period1: Optional start time for charging period 1 (HH:MM format, e.g., "06:00"). charge_end_time_period1: Optional end time for charging period 1 (HH:MM format, e.g., "18:00"). discharge_start_time_period1: Optional start time for discharge period 1 (HH:MM format). discharge_end_time_period1: Optional end time for discharge period 1 (HH:MM format). enable_time_period2: Enable second time period (0=disabled, 1=enabled). Default: 0. charge_start_time_period2: Optional start time for charging period 2 (HH:MM format). charge_end_time_period2: Optional end time for charging period 2 (HH:MM format). discharge_start_time_period2: Optional start time for discharge period 2 (HH:MM format). discharge_end_time_period2: Optional end time for discharge period 2 (HH:MM format).

Returns: API response confirming the command was sent to the inverter.

Raises: ToolError: if the inverter is offline, credentials are invalid, or the API fails.

ParametersJSON Schema
NameRequiredDescriptionDefault
min_socNo
device_snNo
charge_upper_socNo
enable_time_period2No
charge_end_time_period1No
charge_end_time_period2No
charge_from_grid_enableNo
charge_start_time_period1No
charge_start_time_period2No
discharge_end_time_period1No
discharge_end_time_period2No
discharge_start_time_period1No
discharge_start_time_period2No

TDQS

A4.5/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full burden. It discloses that the tool sends a command to the inverter, returns an API response, and raises ToolError for offline inverter, invalid credentials, or API failure. It does not detail potential side effects on other settings, but overall transparency is good.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is structured with a one-line summary, explanatory paragraph, example scenarios, and a clear list of arguments. It is slightly verbose but front-loads the purpose. No unnecessary sentences, but could be trimmed slightly.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given 13 parameters, no output schema, and no annotations, the description covers all aspects: parameter details, error conditions, usage scenarios. It is comprehensive enough for an agent to invoke the tool correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so description must compensate. It provides thorough details for all 13 parameters: defaults, ranges (e.g., min_soc [10,100]), formats (HH:MM for time periods), and explanations of each parameter's purpose. This fully compensates for the lack of schema descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool sets inverter battery to Self Use Mode with configurable thresholds. It uses a specific verb ('Set') and resource ('inverter battery'), and distinguishes from sibling tool get_realtime_data which is read-only.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explains when Self Use Mode is ideal (automated control based on conditions) and provides example automation scenarios. It does not explicitly exclude other modes or provide alternatives, but given the sibling tool, context is clear. Score 4 for clear context without explicit exclusions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updatesv0.1.0
    • First observedget_realtime_data
    • First observedset_battery_self_use_mode

TDQS

A4.3/5.0

Scored across 2 tools

Disambiguation5/5

The two tools have completely distinct purposes: one retrieves real-time data, the other sets a battery mode. There is no ambiguity between them.

Naming Consistency5/5

Both tools follow a consistent verb_noun snake_case pattern (get_realtime_data, set_battery_self_use_mode), making them predictable.

Tool Count2/5

With only 2 tools, the server feels too thin. A typical solar inverter API would require more tools (e.g., historical data, other modes, configuration) to be useful.

Completeness2/5

Major gaps exist: no tool for reading current battery mode, no historical data, no other inverter settings. The set covers only real-time data and one battery mode.

Maintenance

ActivitySlowing
ResponsivenessNo issues

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    D
    maintenance
    Provides access to Alpha ESS solar inverter and battery system data, enabling monitoring of energy statistics, real-time power data, and configuration of battery charging and discharge schedules through the Alpha ESS Open API.
    10
    MIT
  • F
    license
    A
    quality
    Not graded
    maintenance
    Enables monitoring and control of EcoFlow Power Stations and devices through the Model Context Protocol. Users can manage battery levels, toggle AC/DC outputs, and configure charging settings across the DELTA and RIVER series via the EcoFlow API.
    9
    -
  • A
    license
    A
    quality
    D
    maintenance
    Enables real-time solar data from Fronius inverters via Claude, allowing natural language queries about solar production, battery, and grid exchange.
    5
    1
    Apache 2.0
  • A
    license
    A
    quality
    C
    maintenance
    Enables access to Fronius solar inverter data via the MCP protocol, allowing real-time monitoring of energy production, consumption, and battery storage through natural language.
    14
    8
    4
    MIT