Scraping & transparency policy

Who

CouncilTracer is a public-records transparency tool built by Darren Kruse (Kruse Legal). Darren is also a sitting councillor at the City of Mitcham — but CouncilTracer never uses that position. It touches only what's already publicly accessible on a council's website, the same as any resident with a browser. No councillor access, no staff access, no privileged internal system, ever.

What we take

Only documents City of Mitcham already publishes, or is required by law to publish:

• Agendas and minutes for council meetings
• Confidentiality registers — the public record of items held under s90/s91 of the Local Government Act 1999 (SA), and when they're released
• Resolution-tracking reports — outstanding and completed resolutions, where published
• Standing documents — policies, procedures, codes, and by-laws published on the council's own site

Nothing here is obtained through any access a member of the public couldn't also get by visiting the council's site directly.

Why

Council decisions play out across many meetings, over months or years, in documents that are individually public but collectively hard to connect. CouncilTracer reconstructs that history — what was decided, when, what happened to it, and what's still outstanding — so residents, councillors, and journalists don't have to manually piece it together from dozens of separate PDFs. That's the whole product: history before chat, provenance before conclusions.

How we try to be a good citizen about it

• We identify ourselves.Every request carries an honest, self-identifying User-Agent pointing back to this page — we don't disguise scraper traffic as an anonymous browser.
• We rate-limit. Batch operations space requests roughly half a second apart, deliberately, even when we could go faster.
• We check for new documents, we don't hammer for them. Checks run on a schedule matched to how often those documents actually change, not continuously.
• We use a real browser only where genuinely necessary— some council pages are behind bot-management that blocks plain HTTP requests entirely. Where that's the case, we use an ordinary headless browser, at the same modest, scheduled cadence as everything else — not to evade detection, just because a real browser is what's needed to load the page at all, same as it would be for a human visitor.

If you ask CouncilTracer a question

Signing in requires a Microsoft, Google, or Facebook account. We record your account email and sign-in activity, and we keep a count of how many questions each account has asked, to apply fair-use daily limits.

During the current invite-only beta we do store the text of the questions you ask and the answers you receive, along with which sources each answer cited. This is disclosed next to the ask box, and it exists for one reason: so we can check answer quality and find gaps in the corpus while the service is being tested. Please don't include personal information in your questions. If you rate an answer with a thumbs up or down, that rating and any comment you add are stored too, tied to the answer they describe.

Before any wider (non-invite) release, an automated personal-information filter will run before anything is written to storage, and what was collected during the beta will be reviewed. Full detail — what's collected, why, who it's shared with, and how to have it corrected or deleted — is in our privacy policy.

Questions or concerns

If anything here doesn't sit right with you or your council, email darren@kruselegal.com.au. Genuinely — this project exists to make public information more usable, not to cause anyone a headache, and we'd rather hear about a concern directly than have it become friction.