Provides read-only tools to retrieve live power flow, per-circuit consumption, circuit names, daily energy totals, and long-term energy/cost history from a Panasonic AiSEG2 home energy management (HEMS) controller's local web interface.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@aiseg2-mcp what's my current power flow and solar generation?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
日本語版は README.ja.md をご覧ください。
An unofficial, read-only Model Context Protocol server for the Panasonic AiSEG2 home energy management (HEMS) controller. It lets an MCP client (e.g. Claude) read your home's live power flow, per-circuit consumption, circuit names, and daily energy totals from the AiSEG2's local web interface.
This project is not affiliated with or endorsed by Panasonic. "AiSEG" is a Panasonic trademark.
Developed and tested against:
AiSEG2 model MKN713 series
Firmware Ver.2.97I-01
The AiSEG2 web interface is undocumented and changes between firmware revisions. On a different model or firmware the pages this server scrapes may differ and some tools may not work. If you hit a parse error, please open an issue with your model / firmware version.
Related MCP server: emoncms-mcp-server
All tools are read-only (annotated readOnlyHint, non-destructive). The server only issues
GETs and the display-only refresh POSTs the web UI itself uses; it never touches settings or any
/action/ endpoint.
Tool | Returns |
| Instantaneous generation/consumption (kW), buy/sell state, battery status, generation sources, top consuming circuits |
| Every measured circuit's instantaneous draw (W), ranked highest first, with the total |
| Registered circuit ids and names (the authoritative naming source) |
| Today's cumulative generation / consumption / grid-buy / grid-sell (kWh) |
| Long-term energy history from the SD-card export (Wh), long-form points. Args: |
| Long-term energy-cost history from the SD-card export (JPY). Args: |
The two history tools require an SD card inserted in the AiSEG2 — they read the device's SD-card CSV export. The export is downloaded once and cached (see
AISEG_CACHE_DIR/AISEG_CACHE_TTL), so the first call is slow and later calls are fast.
Three ways to run it, depending on your setup.
The simplest option for a local (stdio) MCP client. Requires uv.
AISEG_URL=http://192.168.0.216 AISEG_PASSWORD=... uvx aiseg2-mcpAdd it to Claude Code:
claude mcp add aiseg2 \
--env AISEG_URL=http://192.168.0.216 \
--env AISEG_PASSWORD=your-digest-password \
-- uvx aiseg2-mcpThe container defaults to the streamable-http transport (long-lived network service). Only expose
it behind an authenticating proxy — see Security.
docker run --rm -p 8000:8000 \
-e AISEG_URL=http://192.168.0.216 \
-e AISEG_PASSWORD=your-digest-password \
ghcr.io/chanyou0311/aiseg2-mcp:latestRequires Python 3.12+ and uv.
uv sync
AISEG_URL=http://192.168.0.216 AISEG_PASSWORD=... uv run aiseg2-mcpTo reach the server from claude.ai while your AiSEG2 stays on your LAN, see
examples/remote/ — a Docker Compose stack (MCP + GitHub-OAuth proxy +
Cloudflare Tunnel).
Variable | Required | Default | Description |
| yes | — | AiSEG2 base URL, e.g. |
| yes | — | HTTP Digest password for the AiSEG2 web UI |
| no |
| HTTP Digest user |
| no |
|
|
| no |
| Bind host (streamable-http only) |
| no |
| Bind port (streamable-http only) |
| no |
| Disable the SDK Host allowlist — only behind a trusted auth proxy |
| no |
| Where the SD-card history export is cached |
| no |
| Seconds to reuse a cached history export before re-downloading |
| no |
| Log level |
LAN-only by design. The AiSEG2 speaks plain HTTP with Digest auth; keep it and this server on a trusted local network. The password is read from the environment and is never logged.
Read-only. There is no tool that changes a device setting. The tool surface is enforced by
tests (registered-tool allowlist, tool-name guard, a source scan for /action/, and read-only
annotation checks).
Do not expose the streamable-http transport to untrusted networks without authentication.
This server carries no auth of its own; if you run it as a network service, put an authenticating
reverse proxy in front of it. AISEG_DISABLE_DNS_REBINDING_PROTECTION=true is only appropriate in
that proxied setup.
The AiSEG2 web interface is undocumented; this project builds on the reverse-engineering knowledge shared by prior work:
shimosyan/aiseg2-influxdb-forwarder — the circuit-paging "repeat the last page" terminator and the electric-flow fields.
hiroaki0923/aiseg2-bridge — endpoint and page structure.
Bugfire/aiseg_download — Digest auth and data-endpoint conventions.
Read-only. Get the instantaneous power draw of every measured circuit, ranked highest first.
Pages through the AiSEG2's circuit list and returns each circuit's rank, name and watts, plus
the total measured watts and how many device pages were read. Circuit names here are
display-derived (they may wrap); list_circuits() is the authoritative name source.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| circuits | Yes | |
| page_count | Yes | |
| total_watt | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare read-only, idempotent, and non-destructive. The description adds behavioral detail beyond annotations: it pages through the AiSEG2's circuit list, returns ranks, names, watts, total watts, and page count, and warns that names may wrap. This is useful context not present in structured fields.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences, front-loaded with the core purpose, followed by paging behavior and a name caveat. Every sentence earns its place; no redundancy or fluff.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool has no parameters and an output schema exists, the description covers the necessary context: what it does, how it pages, what it returns, and a name accuracy caveat. This is complete for an agent to select and invoke the tool correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
With zero parameters, the baseline is 4. The description does not need to explain parameters; instead it clarifies what the output contains (rank, name, watts, total, pages), providing semantic value beyond the empty schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states the tool 'Get[s] the instantaneous power draw of every measured circuit, ranked highest first,' which is a specific verb+resource+scope. It distinguishes from the sibling list_circuits by noting that names here are display-derived and list_circuits is authoritative.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides clear context that this is for measuring instantaneous power per circuit and explicitly says list_circuits() is the authoritative name source, guiding when to use the alternative. It does not explicitly enumerate all exclusions but gives practical guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read-only. Query the AiSEG2's long-term energy-cost history from the SD-card export (JPY).
Requires an SD card inserted in the AiSEG2. Shares the same cached download as get_history.
Returns long-form cost points ({timestamp, metric, value}) with values in Japanese yen.
Args:
granularity: "day" takes start/end as YYYY-MM-DD; "month" as YYYY-MM; "year" as YYYY.
start: Range start (inclusive), formatted per the granularity.
end: Range end (inclusive), formatted per the granularity.
limit: Maximum number of series points to return (default 200).
offset: Number of points to skip (for pagination).
| Name | Required | Description | Default |
|---|---|---|---|
| end | Yes | ||
| limit | No | ||
| start | Yes | ||
| offset | No | ||
| granularity | Yes |
| Name | Required | Description |
|---|---|---|
| unit | Yes | |
| series | Yes | |
| has_more | Yes | |
| total_rows | Yes | |
| granularity | Yes | |
| next_offset | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond annotations (readOnly, idempotent), adds unique behavioral details: SD card requirement, cached download sharing with get_history, and return format with JPY values. No contradictions with annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Efficiently structured with purpose first, then requirements, then return format, then args. Every sentence adds value; no fluff.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given output schema and annotations, the description covers purpose, dependencies, return shape, and all parameters thoroughly. No significant gaps for an agent to misuse.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema has no descriptions, so the Args section compensates fully: explains granularity formats, inclusive ranges, limit default, and offset pagination. Adds meaning beyond the raw schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States specific purpose: querying long-term energy-cost history from SD-card export. Distinguishes from sibling get_history by noting cache sharing and cost-specific focus.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides clear context: requires SD card, read-only, shares cached download with get_history. Does not explicitly name exclusions but implies usage for cost history queries.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read-only. Get today's cumulative energy totals (kWh) as of the AiSEG2's current day.
Returns generation, consumption, grid-buy and grid-sell totals for the day. Any meter the device reports as unavailable ("-") comes back as null.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| date | Yes | |
| buy_kwh | No | |
| sell_kwh | No | |
| generation_kwh | No | |
| consumption_kwh | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond the readOnlyHint, idempotentHint, and destructiveHint annotations, the description adds valuable behavioral context: it specifies that unavailable meters are represented as null, and explicitly lists the four totals returned. This informs the agent about response semantics, which is not captured in annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise (three sentences), front-loads the core purpose and read-only nature, and each sentence adds essential information: what it does, what it returns, and null handling. There is no fluff or redundancy with the schema.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple, no-parameter, read-only tool with an output schema, the description is complete. It covers the operation, scope, return values, and edge-case behavior (null for unavailable meters). No additional context is needed, and the output schema handles return format details.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters, so the input schema is trivially covered at 100%. The description correctly does not attempt to explain parameters. The baseline for 0 params is 4, and the description adequately conveys that no arguments are needed.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose with a specific verb ('Get'), resource ('today's cumulative energy totals'), and unit (kWh), and distinguishes it from siblings by focusing on daily cumulative totals. It also enumerates the returned components (generation, consumption, grid-buy, grid-sell), making the scope unambiguous.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage by stating it returns today's totals, but it does not provide explicit guidance on when to use this tool over siblings like get_history or get_power_flow. No alternatives or exclusions are mentioned, so usage is only inferred rather than directly instructed.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read-only. Query the AiSEG2's long-term energy history from its SD-card export (values in Wh).
Requires an SD card inserted in the AiSEG2. The export is downloaded once and cached, so the
first call is slow and later calls are fast. Returns long-form points ({timestamp, metric,
value}); use limit/offset to page (limit caps the number of returned points).
Args:
granularity: Time resolution. "30min"/"hour"/"day" take start/end as YYYY-MM-DD; "month"
takes YYYY-MM; "year" takes YYYY.
start: Range start (inclusive), formatted per the granularity.
end: Range end (inclusive), formatted per the granularity.
metrics: Optional filter by standard series keys (e.g. "generation_pv1", "grid_buy",
"grid_sell", "battery_charge", "battery_discharge", "ev_charge") or their Japanese
header names. Omit for all series.
circuits: Optional filter by circuit name (see list_circuits). Omit for all circuits.
limit: Maximum number of series points to return (default 200).
offset: Number of points to skip (for pagination).
| Name | Required | Description | Default |
|---|---|---|---|
| end | Yes | ||
| limit | No | ||
| start | Yes | ||
| offset | No | ||
| metrics | No | ||
| circuits | No | ||
| granularity | Yes |
| Name | Required | Description |
|---|---|---|
| unit | Yes | |
| series | Yes | |
| has_more | Yes | |
| total_rows | Yes | |
| granularity | Yes | |
| next_offset | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond the annotations (readOnlyHint, idempotentHint), the description discloses valuable behavioral details: the SD-card requirement, caching behavior with first call latency, return format as long-form points, and limit/offset pagination semantics. This significantly informs the agent.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured with a summary sentence, requirements, behavior notes, and an Args list. Every sentence provides useful information without repetition or fluff, making it efficient despite its length.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description covers all aspects needed to use the tool: purpose, prerequisites, caching behavior, return format, and full parameter semantics. Given the output schema exists, return values need no further explanation, making the description complete.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
With 0% schema description coverage, the description entirely compensates by explaining every parameter in detail: granularity-specific date formats, standard metric keys, circuit filtering, and limit/offset defaults. This adds meaning far beyond the bare schema types.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool queries the AiSEG2's long-term energy history from its SD-card export, with a specific verb and resource. It does not explicitly differentiate from sibling tools like get_daily_totals or get_cost_history, so it misses the top score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides clear context for when to use the tool: it requires an SD card, notes the first call is slow due to caching, and explains pagination. However, it does not mention alternatives or when not to use it, so it lacks explicit exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read-only. Get the AiSEG2's instantaneous whole-home power flow right now.
Returns current generation and consumption (kW), whether the home is buying or selling grid power, storage-battery status if a battery is connected, the per-source generation breakdown, and the top consuming circuits at this moment.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| battery | No | |
| buy_sell | Yes | |
| generation_kw | Yes | |
| top_consumers | No | |
| consumption_kw | Yes | |
| generation_detail | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already indicate read-only, idempotent, non-destructive behavior, so the description adds value by detailing what data is returned and under what conditions (e.g., 'if a battery is connected'). It goes beyond the annotations by specifying the return contents, although it does not discuss error cases or access requirements.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences long, front-loads the read-only and time-sensitive nature, and packs useful detail into a compact list of return categories. Every sentence earns its place, with no redundancy or filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity, zero parameters, and existing output schema, the description fully covers what an agent needs: it states the resource, the temporal scope, the return categories, and conditional elements. No critical invocation details are missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters and the schema coverage is 100% (empty properties), so there is no parameter meaning to clarify. The baseline for 0-parameter tools is 4, and the description appropriately focuses on output rather than input semantics.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb ('Get'), identifies the exact resource ('AiSEG2's instantaneous whole-home power flow'), and clearly scopes it to 'right now.' It enumerates distinct return values (generation/consumption, buy/sell status, battery, per-source breakdown, top circuits) that distinguish it from the historical/circuit-focused sibling tools.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly conveys real-time, current-state usage through words like 'instantaneous' and 'right now,' which implies use for immediate monitoring rather than historical analysis. It does not explicitly name alternatives like get_history or get_circuit_breakdown, but the context is clear enough for an agent to select this tool for a live snapshot.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read-only. List the registered measurement circuits with their stable ids and names.
This is the AUTHORITATIVE source of circuit naming (from the device's installation settings).
The names in get_circuit_breakdown() are display-derived and may differ (line wraps); prefer
these when you need a canonical circuit name.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| circuits | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Adds context beyond annotations: authoritative source from installation settings, notes that names may differ due to line wraps. Annotations already cover read-only/idempotent, so description adds provenance.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences, front-loaded with read-only and list action, then essential nuance. Zero waste.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a zero-param tool with output schema and strong annotations, the description fully explains purpose, authoritative nature, and relationship to a sibling. Complete.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters; schema trivially covers everything. Baseline 4 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Specific verb+resource: lists registered measurement circuits with stable ids and names. Distinguishes from sibling get_circuit_breakdown by noting authoritative naming.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicitly states when to prefer this tool (canonical circuit names) and cites get_circuit_breakdown as the alternative with display-derived names. Clear usage context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.1.0get_circuit_breakdownget_cost_historyget_daily_totalsget_historyget_power_flowlist_circuitsScored across 6 tools
Each tool targets a distinct aspect: instantaneous overview, per-circuit draw, canonical names, daily totals, historical energy, and historical cost. The only slight overlap is get_power_flow's top circuits vs get_circuit_breakdown, but descriptions clearly differentiate summary versus full detail.
All tools use snake_case verb_noun patterns: get_ for data retrieval and list_ for enumerating resources. The naming is consistent and predictable, following a clear convention.
6 tools is well-scoped for a read-only energy monitoring API. Each tool provides a distinct piece of data without being redundant or overwhelming.
The tool set covers current status, per-circuit breakdown, canonical circuit list, daily cumulative totals, long-term history, and cost history. For a monitoring-focused server, this is complete and leaves no obvious dead ends.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
MCP server wrapping the Tesla Fleet API and TeslaMate API
An MCP server that provides read access to your cloud storage providers, bank accounts and more.
MCP server for Withings health data — sleep, activity, heart, and body metrics.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/chanyou0311/aiseg2-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
