Endpoint
https://camfeed.cameronwilson.com.au/mcp
The MCP endpoint exposes the same public items as RSS and JSON, but with collection-level resources, item-level records, and read-only discovery tools for AI clients.
Search without an AI connector
Search published work, or request /search.json?q=search+terms. Results include snippets and links to complete HTML, JSON and Markdown.
/archive.json returns at most 200 records per page with coverage, total_count and next_url. Follow next_url until null. The archive covers selected work from 2026-05-10T00:00:00+10:00.
Connect an AI client
Use the Streamable HTTP endpoint below. No account, API key, or OAuth flow is required.
https://camfeed.cameronwilson.com.au/mcpChatGPT: enable Developer mode, add a new MCP app/connector, and paste the endpoint URL.
Claude Code:
claude mcp add --transport http camfeed https://camfeed.cameronwilson.com.au/mcpCodex:
codex mcp add camfeed --url https://camfeed.cameronwilson.com.au/mcpCursor and JSON-configured clients:
{"mcpServers":{"camfeed":{"url":"https://camfeed.cameronwilson.com.au/mcp"}}}Public Resources
camfeed://about, camfeed://press-kit, camfeed://reporting-guide, camfeed://beats, camfeed://coverage-profile, camfeed://latest, camfeed://archive, camfeed://featured, camfeed://topics, camfeed://appearances, camfeed://media, camfeed://commentary, camfeed://citation-guide, camfeed://items/{id}, and camfeed://stories/{id}.
Question prompts
ask_about_camerons_reporting answers source-facing questions from Cameron's published reporting guide. It links the canonical guide, treats attribution terms as Cameron's stated practices rather than universal rules, and directs sensitive disclosures to Cameron instead of an AI client.
Read-Only Tools
search and fetch provide the OpenAI-compatible knowledge-retrieval contract. Richer archive tools are latest_work, featured_work, search_journalism (optional query, with kind/source/from/to filters and matched snippets), get_item (with related_items), work_by_topic, and recommended_reading.