For AI agents
ViewJobs is regional Australia's jobs platform. Every live ad has an indexable page with schema.org JobPosting markup, and assistants can search the whole catalogue through a free, read-only MCP server. Applying always happens on the ad page; nothing here applies on a person's behalf.
Search from an assistant
The MCP endpoint is being prepared. This page will carry the address and setup steps when it opens.
Tools
| Tool | What it does |
|---|---|
| search_jobs | Search live job ads across regional and metropolitan Australia by keywords, optionally near a place. Returns compact cards (title, employer, location, work type, pay, posted date, apply type, URL). Give the person the card's `url` — it is the page to read and apply from. Use `get_job` for the full description. |
| get_job | Fetch one job ad by its id (from `search_jobs`): the full description as plain text, plus three similar live ads. An expired ad is reported as expired with its URL, not dropped. |
| list_roles | The most-advertised roles right now, nationally or for a town or region, with live ad counts and the hub page for each (a browsable list of that role's ads). Use it to discover what is hiring somewhere before searching. |
| list_places | Resolve a place name to the towns and regions ViewJobs knows, with live ad counts and each place's hub page. Use it when a person names somewhere ambiguous (there are several Richmonds) or to see how much is hiring nearby. |
| salary_bands | Advertised pay bands (25th, median, 75th percentile, annual AUD) for a role, computed from published job ads. National only. Says how many ads the band rests on and when it was computed; do not present it as an offer. |
| search | ChatGPT deep-research alias of search_jobs: a plain query, up to 10 results with id, title and url. Prefer search_jobs for filters. |
| fetch | ChatGPT deep-research alias of get_job: one ad by id as { id, title, text, url, metadata }. |
Limits and attribution
- Read-only. At most 20 results per page, pages 1–50; descriptions are capped at 4,000 characters.
- Abusive traffic may be blocked. If you need sustained high volume for a legitimate use, contact support@viewjobs.com.au.
- Every link the server returns carries
utm_medium=ai_agentand the client name, so we can see which assistants send people to jobs. We log each call's tool, arguments (the query, place or ad id), a salted hash of the address and the latency for 90 days; nothing identifies a person. - Featured ads appear in results as ordinary listings marked
featured: true.
Crawlers
Search-index and user-fetch agents (OAI-SearchBot, Claude-SearchBot, Claude-User, ChatGPT-User, PerplexityBot, Perplexity-User, Googlebot, Bingbot, Applebot and the rest) may read every public page and cite it. Training crawlers may read hub and marketing pages but not individual job ads. The full policy is in robots.txt; the machine-readable index is llms.txt.
