aemo-mcp

Ask Claude about Australia's electricity market and get real, current numbers — 5-minute spot prices, regional demand, generation by fuel, interconnector flows, rooftop solar — not "I don't have access to that data." This MCP server gives Claude (and other MCP clients like Cursor) live access to the Australian Energy Market Operator (AEMO) NEMWEB feeds, with curated mappings for the most-asked indicators.

Companion to abs-mcp (ABS macro stats), rba-mcp (RBA interest + FX rates), ato-mcp (ATO tax + ACNC charity), apra-mcp (banking + superannuation), aihw-mcp (health & welfare), asic-mcp (companies + financial advisers), and au-weather-mcp (Australian weather) — together they cover the most-asked Australian official data.

What you can ask

Once installed, your LLM can answer questions like:

Question	What the tool does
What's the current NSW spot price?	latest("dispatch_price", filters={"region":"NSW1"})
Did SA hit negative pricing in the last 24 hours?	get_data("dispatch_price", filters={"region":"SA1"}, start_period=…) and filter value < 0
Generation by fuel type right now in QLD	latest("generation_scada", filters={"region":"QLD1"}), aggregated by fuel
Weekly average dispatch price for VIC, last 4 weeks	get_data("dispatch_price", filters={"region":"VIC1"}, start_period=…)
Rooftop PV forecast for tomorrow	get_data("rooftop_pv", filters={"section":"forecast"}, start_period="<tomorrow>")
What's the current flow across Heywood (VIC ↔ SA)?	latest("interconnector_flows", filters={"interconnector":"V-SA"})
Total NEM demand right now	latest("dispatch_region")

Every answer comes with the interval timestamp (AEMO market time, UTC+10), units (MW, $/MWh), and a link back to the NEMWEB source. The MCP wraps NEMWEB's CSV/ZIP feeds and exposes them through 5 plain-English tools.

Install

# After publish:
uvx --upgrade aemo-mcp

# Local dev:
uv pip install -e .

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "aemo": {
      "command": "uvx",
      "args": ["--upgrade", "aemo-mcp"]
    }
  }
}

Why --upgrade? uvx aemo-mcp (without the flag) uses whatever wheel is cached and never adopts new PyPI releases on its own. --upgrade makes uvx check PyPI on each launch and pull a newer release if one exists. Recommended for everyone except offline-first / pinned-version workflows. To verify which version is currently serving you, look at the server_version field on any DataResponse.

If you also have rba-mcp / abs-mcp installed, all servers run side-by-side. Claude disambiguates with the server prefix (aemo:get_data, rba:get_data, abs:get_data).

For local dev (pre-PyPI):

{
  "mcpServers": {
    "aemo": {
      "command": "uv",
      "args": ["run", "--directory", "/absolute/path/to/aemo-mcp", "aemo-mcp"]
    }
  }
}

Cursor

Add to ~/.cursor/mcp.json (or workspace .cursor/mcp.json):

{
  "mcpServers": {
    "aemo": {
      "command": "uvx",
      "args": ["--upgrade", "aemo-mcp"]
    }
  }
}

Tools

Tool	What it does
search_datasets(query, limit=10)	Fuzzy-search the 7 curated NEM datasets by name, topic, or region.
describe_dataset(dataset_id)	Schema, filters, source URL, cadence, and example queries for one dataset.
get_data(dataset_id, filters, start_period, end_period, format)	Query data. Records / series / csv format.
latest(dataset_id, filters)	Most recent 5-min or 30-min interval for time-series feeds.
list_curated()	The 7 curated dataset IDs.

Curated datasets

The 7 datasets cover ~95% of typical NEM analytic queries:

dataset_id	Cadence	Source	Use case
dispatch_price	5 min	DispatchIS / DISPATCHPRICE	Current spot price per region; negative-pricing detection
dispatch_region	5 min	DispatchIS / DISPATCHREGIONSUM	Demand + scheduled + semi-scheduled gen + net interchange
interconnector_flows	5 min	DispatchIS / DISPATCHINTERCONNECTORRES	MW flow + losses across NEM interconnectors
generation_scada	5 min	Dispatch_SCADA	DUID-level MW (every unit), aggregable by fuel
rooftop_pv	30 min	ROOFTOP_PV/ACTUAL + FORECAST	Regional rooftop solar (actual + forecast)
predispatch_30min	30 min	PredispatchIS	30-min forecast, ~40h horizon
daily_summary	Daily	Daily_Reports	Yesterday's full data in one drop

Use list_curated() to enumerate, describe_dataset(dataset_id) to learn the filters available on each.

Regions

NSW1 (New South Wales), QLD1 (Queensland), SA1 (South Australia), TAS1 (Tasmania), VIC1 (Victoria). Western Australia (WEM) and the Northern Territory are not on the NEM and are out of scope.

Trust contract

Every DataResponse carries:

• source = "Australian Energy Market Operator"
• attribution — AEMO Copyright Permissions verbatim attribution string
• source_url — the NEMWEB folder the data came from
• retrieved_at — UTC timestamp of the fetch
• interval_start / interval_end — period covered
• stale — True if the latest interval is older than 2× the feed cadence
• server_version — the wheel that served the call

Licence + attribution

This package is MIT-licensed (see LICENSE).

The AEMO data it fetches is published under AEMO's Copyright Permissions policy: AEMO grants general permission to use AEMO Material for any purpose (commercial included) on the sole condition of accurate attribution of the relevant material and AEMO as its author. See https://aemo.com.au/privacy-and-legal-notices/copyright-permissions.

End-users redistributing data fetched via this server must credit AEMO. The canonical attribution string is on every DataResponse.attribution.

Worked examples

"What's the current NSW spot price?"

latest(dataset_id="dispatch_price", filters={"region": "NSW1"})

→ {"records": [{"period": "2026-05-14T10:05:00+10:00", "value": 87.5, "dimensions": {"region": "NSW1", "metric": "rrp"}, "unit": "$/MWh"}], ...}

"NSW spot price for the last 24 hours"

get_data(dataset_id="dispatch_price", filters={"region": "NSW1"},
         start_period="2026-05-13", end_period="2026-05-14")

"Did SA hit negative pricing in the last 24 hours?"

get_data(dataset_id="dispatch_price", filters={"region": "SA1"},
         start_period="<24h ago>")

Then the LLM filters value < 0 client-side.

"Generation by fuel type right now in QLD"

latest(dataset_id="generation_scada", filters={"region": "QLD1"})

→ DUID-level rows with fuel attribution; the LLM aggregates by fuel.

How it works

• Live-fetch only. No NEMWEB archives in the wheel. Every request goes through the cache.
• Cache TTLs tuned per cadence. 60s for 5-min feeds, 5min for 30-min feeds, 1h forecasts, 24h daily archive. Timestamped historical files are immutable in NEMWEB and cache effectively forever.
• In-flight request deduplication. Concurrent callers for the same URL share one HTTP request. Critical at 5-min cadence with many users.
• Latest-file detection is purely lexicographic on the NEMWEB directory listing — AEMO embeds the interval timestamp in every filename (PUBLIC_DISPATCHIS_YYYYMMDDHHmm_<seq>.zip), so max() is enough.
• Multi-section CSV parser handles AEMO's I,/D, row format where a single ZIP holds several tables (DISPATCHPRICE, DISPATCHREGIONSUM, DISPATCHINTERCONNECTORRES, etc.).

Development

git clone https://github.com/Bigred97/aemo-mcp
cd aemo-mcp
uv sync --extra dev
uv pip install -e .
uv run pytest -q

Run live tests (hit NEMWEB):

uv run pytest -q -m live

Zero-flake validation:

for i in $(seq 1 10); do uv run pytest -q || break; done

Sister MCPs (Australian Public Data portfolio)

• abs-mcp — Australian Bureau of Statistics (CPI, unemployment, ERP, building approvals)
• rba-mcp — Reserve Bank of Australia (cash rate, lending stats, exchange rates)
• ato-mcp — Australian Taxation Office (tax stats, ACNC charities)
• apra-mcp — Australian Prudential Regulation Authority (banking, insurance, super)
• aihw-mcp — Australian Institute of Health and Welfare
• asic-mcp — Australian Securities and Investments Commission (company registers)
• aemo-mcp — this one. Australian Energy Market Operator (NEM dispatch, spot prices, generation).
• au-weather-mcp — Open-Meteo (Bureau of Meteorology aggregator)
• wgea-mcp — Workplace Gender Equality Agency
• aus-identity — Postcode / state / ABN normalisation helper used by all sisters

Author

Built by Harry Vass. Issues + PRs welcome at github.com/Bigred97/aemo-mcp.