Skip to main content
Glama
Australian Prudential Regulation Authority

Australian Prudential Regulation Authority

apra-mcp

mcp-name: io.ausdata/apra-mcp

PyPI Python License Tests CodeQL Glama MCP server quality

MCP server for Australian Prudential Regulation Authority statistics. Plain-English access to per-bank capital ratios, fund-by-fund superannuation, and post-AASB17 life + general insurance — every prudentially-regulated entity in Australia, every quarter, from a single uvx command.

Hosted access? For cross-source queries, webhooks, an always-on REST API, and a uniform response envelope across all 9 sources, see ausdata.io — free tier available (500 calls/mo, no card).

"What's CBA's CET1 ratio?"
"Which super fund has the most members?"
"Top 10 banks by total capital, latest quarter"
"Gross written premium for the general insurance industry, post-AASB17"
"Largest life insurance product groups by claims"

Sister to abs-mcp, rba-mcp, ato-mcp, and au-weather-mcp.


Install

uvx --upgrade apra-mcp

Claude Desktop

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

Claude Code

claude mcp add apra --command uvx --args -- --upgrade apra-mcp

Related MCP server: ato-mcp

What it exposes

Six tools, all plain-English in, structured out:

Tool

Purpose

search_datasets

Fuzzy-search the curated catalog by keyword

describe_dataset

List a dataset's filterable dimensions and returnable measures

get_data

Query with filters, measures, period range, output format

latest

Last observation per measure (shortcut)

top_n

Rank rows by a measure, return top (or bottom) N

list_curated

Enumerate the curated dataset IDs

Every response is the same shape — dataset_id, dataset_name, query, period, unit, row_count, records, apra_url, download_url, framework (insurance only), attribution, stale flag, server_version — across every curated dataset.


Curated datasets (7 in v0.1)

ID

What it is

Period

ADI_KEY_STATS

Per-bank CET1 / Tier 1 / Total capital + RWA

latest quarter snapshot

ADI_RISK_WEIGHTED_ASSETS

Per-bank RWA by risk type (credit / operational / market / IRRBB)

latest quarter snapshot

SUPER_FUND_LEVEL

Fund-by-fund members, benefits, demographics

latest quarter snapshot

INSURANCE_GENERAL

Long-format general insurance (post-AASB17, ~24k rows × 10 quarters)

Sep 2023 → latest

INSURANCE_GENERAL_HISTORICAL

General insurance archive (pre-AASB17)

Dec 2002 → Jun 2023

LIFE_INSURANCE

Long-format life insurance (post-AASB17, ~10k rows × 10 quarters)

Sep 2023 → latest

LIFE_INSURANCE_HISTORICAL

Life insurance archive (pre-AASB17)

Jun 2008 → Jun 2023

Snapshot vs time-series. ADI and Super datasets ship the most recent reporting quarter only (APRA refreshes the file each quarter). The four insurance datasets are long time series in a single file. Pass start_period / end_period as ISO dates (2025-12-31), bare years (2024), year-months (2025-06), or quarter shorthand (2025-Q4) — all normalised internally.

Cross-source compatibility. The state_territory filter on INSURANCE_GENERAL accepts canonical state codes ("NSW"), full names ("New South Wales"), case-insensitive variants ("nsw"), ISO 3166-2 ("AU-NSW"), and 4-digit postcodes ("2000" → New South Wales). Powered by aus-identity — the same input format works across abs-mcp, ato-mcp, apra-mcp, aihw-mcp, and asic-mcp.


Reliability — 3-tier URL resolution

APRA publishes XLSX at date-versioned paths that change every quarter. apra-mcp resolves them through three tiers:

  1. Live scrape — fetch the canonical APRA landing page (with ETag conditional-GET so refreshes between releases cost zero bytes), regex-extract the .xlsx href matching the dataset's filename pattern, pick the latest-dated match. Cached 6h.

  2. Bundled seed manifest — when the live scrape fails, fall back to data/seed_urls.json shipped in the wheel. CI refreshes the manifest daily. The response is flagged stale: true with an honest reason.

  3. YAML default — last-resort URL from the curated YAML.

Net effect: a fresh uvx apra-mcp always gets the current quarter; a 3-month-old install still works because the seed manifest is refreshed and --upgrade pulls a new wheel.


Framework break (insurance only)

APRA changed the reporting framework on 1 July 2023 (AASB 17 Insurance Contracts + capital framework revision). Pre- and post-break data are not directly comparable — APRA's own guidance is explicit. apra-mcp ships paired datasets:

  • INSURANCE_GENERAL (post-AASB17) + INSURANCE_GENERAL_HISTORICAL (pre-AASB17)

  • LIFE_INSURANCE (post-AASB17) + LIFE_INSURANCE_HISTORICAL (pre-AASB17)

Every response on an insurance dataset includes a framework block surfacing the break + a historical_dataset cross-reference, so agents see the warning before splicing series.


Attribution

Data sourced from the Australian Prudential Regulation Authority. Licensed under Creative Commons Attribution 3.0 Australia (CC BY 3.0 AU). apra-mcp is MIT-licensed; APRA's data carries the upstream CC-BY 3.0 AU licence, echoed in every response's attribution field.


Sister MCPs (Australian Public Data portfolio)

Want all 9 sources behind one REST API? The hosted gateway at ausdata.io adds cross-source joins, full history, webhooks, and HMAC-signed responses on top of these MCPs — free tier (500 calls/mo, no card).

  • 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 — this one. Banks, super, insurance.

  • aihw-mcp — Australian Institute of Health and Welfare

  • asic-mcp — Australian Securities and Investments Commission (company registers)

  • aemo-mcp — 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


Development

git clone https://github.com/Bigred97/apra-mcp.git
cd apra-mcp
uv venv
uv pip install -e ".[dev]"
pytest                  # unit tests
pytest -m live          # integration tests against apra.gov.au

Issues and contributions welcome: github.com/Bigred97/apra-mcp/issues.

Available Tools

6 tools
describe_datasetA

Describe a dataset's filterable dimensions, returnable measures, units, source, and (for insurance) framework break info.

Use this before calling get_data on a new dataset — it tells you the valid filter keys ('institution', 'sector', 'data_item'), the valid enumerated filter values ('cba', 'major_banks'), the measure aliases ('cet1_ratio', 'total_capital'), and the canonical source URL.

For insurance datasets, the response includes a framework block documenting the Q3-2023 AASB-17 break.

Returns: DatasetDetail with id, name, description, period_coverage, list of dimensions, list of measures, source_url, download_url, and optional framework info.

Inputs: dataset_id (required). Returns: describe_dataset shape.

get_dataA

Query a curated APRA dataset and return observations.

Examples: # CBA's CET1 ratio over time resp = await get_data( "ADI_KEY_STATS", filters={"institution": "cba"}, measures="cet1_ratio", )

# Major banks' total capital, last 5 quarters
resp = await get_data(
    "ADI_KEY_STATS",
    filters={"sector": "major_banks"},
    measures="total_capital",
    start_period="2024-01-01",
)

# Total industry gross written premium (general insurance)
resp = await get_data(
    "INSURANCE_GENERAL",
    filters={"data_item": "Gross written premium",
             "industry_segment": "total_industry"},
)

# AustralianSuper member account counts
resp = await get_data(
    "SUPER_FUND_LEVEL",
    filters={"fund_name": "australian_super"},
    measures=["total_member_accounts", "total_members_benefits"],
)

Returns: DataResponse with records (or csv), unit, period bounds, row_count, source URL, the actual download_url used, optional framework info (insurance only), and CC-BY 3.0 AU attribution.

Inputs: dataset_id (required), format, filters, measures, +2 more. Returns: get_data / latest / top_n shape — uniform across curated datasets. `records` ca…

latestA

Return the most recent observation per measure for a dataset.

Trims to the single latest period per measure across the filtered slice — useful for "what's CBA's current CET1?" style questions without having to think about start_period. Wide-layout entity-level datasets (INSURANCE_GENERAL etc.) are capped at limit rows.

Examples: # Latest CBA capital ratios resp = await latest("ADI_KEY_STATS", filters={"institution": "cba"})

Inputs: dataset_id (required), limit, filters, measures. Returns: get_data / latest / top_n shape — uniform across curated datasets. `records` ca…

list_curatedA

List every curated dataset ID in this version of apra-mcp.

Returns: Sorted list of dataset IDs.

Inputs: no named parameters. Returns: result

search_datasetsA

Fuzzy-search the curated APRA dataset catalog.

All datasets ship hand-curated in v0.1: per-bank capital ratios, per-bank risk-weighted assets, fund-by-fund superannuation, and post-AASB17 life + general insurance (with separate historical archives for the pre-Q3-2023 reporting framework).

Examples: # Find the dataset for bank capital ratios results = await search_datasets("bank capital cet1") # → [{id: 'ADI_KEY_STATS', name: 'ADI Key Statistics — entity-level...', ...}]

# Discover what's available on insurance
results = await search_datasets("insurance premium")

Returns: List of DatasetSummary (id, name, description, update_frequency, is_curated), ranked by relevance.

Inputs: query (required), limit. Returns: result

top_nA

Return the N rows with the largest (or smallest) value of a measure.

The single most common agent workflow: "show me the top 10 X by Y". top_n does the rank server-side and returns only the requested rows.

Examples: # Biggest 10 banks by total capital, latest quarter top_n("ADI_KEY_STATS", "total_capital", n=10, filters={"period": "2025-12-31"})

# Most members per super fund (latest)
top_n("SUPER_FUND_LEVEL", "total_member_accounts", n=10,
      filters={"period": "2025-12-31"})

# 5 lowest CET1 ratios in the latest quarter
top_n("ADI_KEY_STATS", "cet1_ratio", n=5, direction="bottom",
      filters={"period": "2025-12-31"})

Returns: DataResponse with at most n records, sorted by measure in the requested direction. Other fields match get_data.

Inputs: measure (required), dataset_id (required), n, filters, +1 more. Returns: get_data / latest / top_n shape — uniform across curated datasets. `records` ca…

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    A4.2/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool has a distinct and clearly defined purpose: describe_dataset for metadata, get_data for querying, latest for most recent values, list_curated for dataset IDs, search_datasets for discovery, and top_n for ranking. There is no overlap or ambiguity between tool functions.

    Naming Consistency5/5

    All tool names follow a consistent snake_case verb_noun or verb pattern (e.g., describe_dataset, get_data, list_curated, search_datasets). Even 'latest' functions as a verb in context. No mixing of conventions.

    Tool Count5/5

    Six tools is appropriate for the server's purpose of querying curated APRA datasets. It covers discovery, metadata, data retrieval, latest observations, and top ranking without being excessive or thin.

    Completeness5/5

    The tool set provides a complete workflow for interacting with curated datasets: search and list to find datasets, describe to understand structure, get_data for arbitrary queries, latest for current values, and top_n for ranking. No essential operations are missing.

    Maintenance

    ActivityNo data
    ResponsivenessSyncing

    Related MCP Connectors

    Related MCP Servers