Privacy Policy
Last updated: 18 November 2026.
ausecon is an independent, open-source project. It is not operated by, affiliated with, or endorsed by the Australian Government, the Australian Bureau of Statistics (ABS), the Reserve Bank of Australia (RBA), or the Australian Prudential Regulation Authority (APRA). It retrieves data those agencies publish openly.
This policy covers three things:
- the hosted MCP server at
https://ausecon-mcp-server.onrender.com/mcp(also reachable ashttps://mcp.auseconmcp.com/mcp), which is what you use when you add ausecon as a connector in Claude or another client; - the installable package (
ausecon-mcp-serveron PyPI, and its container image), which you run on your own machine; - this documentation site at
auseconmcp.com.
The short version
Section titled “The short version”- There are no accounts, no sign-in, no API keys and no cookies.
- The server does not collect personal information, and does not sell, rent or share anything about you.
- The server keeps no record of who called it. Its logs record which public dataset was fetched, and the error message when a call fails, but not who made the request.
- The only things sent on to the ABS, RBA or APRA are dataset and series identifiers and date ranges.
Hosted MCP server
Section titled “Hosted MCP server”What the server receives
Section titled “What the server receives”When your AI assistant calls an ausecon tool, the server receives the tool name and its arguments.
These are things like a dataset identifier (CPI), a series identifier, a date range, a number of
observations, or a short search phrase such as “cash rate”. Your assistant chooses these arguments.
The server never sees your conversation, your files or your account details.
Because the assistant writes the arguments, a search phrase could in principle contain text from your conversation. The server uses that text only to answer the request, and keeps it only in the error log described under Logging if the call fails.
Like any website, the server also receives standard connection information, such as your client’s IP address and request headers. These are handled by the hosting provider (see below).
What the server does with it
Section titled “What the server does with it”The server uses each request only to answer it:
- Search and discovery tools (
search_datasets,list_catalogue,list_economic_concepts,describe_dataset) match your request against a catalogue built into the server. Nothing is sent anywhere else. - Retrieval tools request the matching public data from the ABS, RBA or APRA. Those requests contain only dataset and series identifiers and date bounds. They come from the server, so the agencies see the server’s address, not yours.
- The release calendar (
list_release_events) is read from fixed public ABS and RBA pages.
Logging
Section titled “Logging”The server writes operational logs so the maintainer can see whether upstream sources are working.
A log line records the event (for example request.success), the source (abs, rba or apra),
the public dataset identifier, the upstream URL, the HTTP status, the duration and the response size.
Calls to the five retrieval tools (get_economic_series, get_derived_series, get_rba_table,
get_abs_data and get_apra_data) also log the tool name, whether the client said it can display
inline charts, and whether the result was charted. That log line
records no arguments and nothing about who called.
The server does not log:
- HTTP access logs (they are turned off);
- client IP addresses or request headers;
- request or response bodies;
- the arguments of successful tool calls, including search phrases.
When a tool call fails, the error is logged so the maintainer can fix the cause. The error message can include the argument that caused it, for example an unrecognised concept name. These error logs are never linked to a client’s IP address or identity.
Caching
Section titled “Caching”To avoid fetching the same public data repeatedly, the server keeps copies of ABS, RBA and APRA
responses in memory and on disk. A copy is treated as fresh for up to an hour. An older copy may be
returned, clearly marked stale, if a source is temporarily unavailable. The cache holds only public
statistical data. It holds nothing about who requested it.
The hosted server’s cache lives on the hosting provider’s temporary storage, which is cleared whenever the service is redeployed or restarted.
Sessions
Section titled “Sessions”The MCP protocol gives each connection a session identifier (mcp-session-id). It is used only to
route your client’s requests and is not linked to any identity. It is not authentication, and it is
held in memory only.
Hosting provider
Section titled “Hosting provider”The hosted server runs on Render. Render handles network traffic to the server, and may keep its own infrastructure and request logs under its own terms; see Render’s privacy policy. The ausecon project does not add tracking, analytics or advertising to the hosted server, and does not export its logs to any other service.
Installable package
Section titled “Installable package”When you install ausecon from PyPI or run its container image yourself, it runs entirely on your machine. It contacts only the public ABS, RBA and APRA endpoints needed to answer your requests. It sends no telemetry or usage data to the ausecon project or anyone else.
Its cache is stored in your user cache directory (~/.cache/ausecon-mcp/ on Linux,
~/Library/Caches/ausecon-mcp/ on macOS). You can disable it with AUSECON_CACHE_DISABLED=1.
Documentation site
Section titled “Documentation site”This site is hosted on Vercel and uses Vercel Web Analytics and Speed Insights. They measure page views and page performance in aggregate, so the maintainer can see which pages are useful and whether the site is fast. Query strings and URL fragments are removed before performance data is sent. See Vercel’s privacy policy for how Vercel processes this data.
Data sharing and retention
Section titled “Data sharing and retention”ausecon does not sell, rent or trade any data. Apart from the hosting providers named above, it shares nothing with third parties. The only data sent to the ABS, RBA or APRA is the dataset identifiers and date ranges described above.
Cached public data is kept for as long as the cache needs it and is cleared on redeploy. Operational logs are kept only for as long as the hosting provider’s log retention for the service.
Children
Section titled “Children”ausecon is a statistical data service. It is not directed at children and does not knowingly collect information from anyone.
Changes to this policy
Section titled “Changes to this policy”If this policy changes, the new version will be published on this page with a new “Last updated” date. The full history is in the project repository.
Contact
Section titled “Contact”For privacy questions or requests, start a discussion at github.com/AnthonyPuggs/ausecon-mcp-server/discussions.
To report a security vulnerability, do not open a public issue. Use the private reporting process in the project’s security policy.
ausecon is an independent open-source project. It is not affiliated with, endorsed by or operated by the Australian Government, the Australian Bureau of Statistics, the Reserve Bank of Australia or the Australian Prudential Regulation Authority. Data is published by those agencies under CC BY 4.0; see Data Sources and Licensing.
