Data for developers and AI agents

JSON API

Read-only, described in OpenAPI at /api/openapi.json.

• GET /api/v1/search?q=preston finds seats, politicians and advertisers by name, each with its page URL.
• GET /api/v1/spenders?jurisdiction=vic&side=labor&limit=10 ranks advertisers by Facebook and Instagram ad spend. Jurisdictions: federal, vic, nsw, qld, wa, sa, tas, act, nt. Sides: labor, liberal, nationals, greens, independent, minor, third_party, union.
• GET /api/v1/politicians/anthony-albanese returns one politician: party, seat, a sourced profile and ad spend.
• GET /api/v1/status says when the data was last updated.

Spend is always a range, {"min": 40000, "max": 57000, "currency": "AUD"}, because Meta publishes each ad's spend as a band. Please quote it as a range.

MCP server

Connect an AI assistant to https://election.computer/mcp (Streamable HTTP, no sign-in). It offers search, top_spenders and get_politician. Connection details are in the server card. In a browser that supports WebMCP, the same tools are offered on every page.

Pages as Markdown

Every page has a Markdown version: request it with Accept: text/markdown, or add .md to the address (/about.md, and /index.md for the home page).

For AI agents

• /llms.txt: what this site is, when to use it and how
• /.well-known/agent-skills/index.json: an agent skill
• /.well-known/api-catalog and /.well-known/ai-catalog.json: machine-readable catalogues
• /auth.md: there is nothing to sign up for

Fair use

The data changes once a week, so cache what you fetch and keep to a few requests a second. Ads and spend come from Meta's Ad Library; seats and members from the parliaments and Wikipedia. See how this works for what the numbers can and can't tell you.