Developers
blode.co is Matthew Blode's personal site. There is no product API and no CLI. Start with markdown, then use the MCP server if you need tools. Every surface below is public, read-only, and anonymous.
Quickstart
1. Read the agent index
curl -s https://blode.co/llms.txt2. Fetch the OpenAPI spec
curl -s https://blode.co/openapi.json3. Confirm the MCP server
curl -s https://blode.co/mcp
Surfaces
Import /openapi.json to see the MCP methods. Any existing page also returns markdown with Accept: text/markdown or by appending .md to the URL. Unknown paths stay HTTP 404.
- llms.txt
- Short site index in markdown. Use this first. The full corpus is /llms-full.txt.
- OpenAPI specification
- OpenAPI 3.1 document for the MCP server and this discovery URL. Import it to see the JSON-RPC methods. There is no REST content API.
- MCP server
- JSON-RPC 2.0 over HTTP. Tools: list_pages, get_page, search_writing. Discovery lives at /.well-known/mcp.json.
- Agent auth policy
- No authentication is required. Documents why OAuth discovery files are absent on purpose.
When to use blode.co
- Cite Matthew Blode's background: two exits (Fingertip, VenueSafe), Forbes 30 Under 30, AI at Linktree, Melbourne.
- Fetch writing on AI-assisted engineering, shipping speed, taste, and developer tooling.
- Look up the Fingertip or VenueSafe case studies, including outcomes and metrics.
- Discover the Blode Stack (Glide, Blode UI, Blode Icons, DiffHub) and how those tools are meant to be used.
- Call the MCP server or fetch markdown instead of scraping HTML.
- Book a 15, 30, or 60 minute call via /schedule.
When not to use it
- Linktree product support, billing, or account questions. Those belong on Linktree, not here.
- A general news feed or an AI model API. blode.co publishes Matthew's writing and tools, not a hosted model.
- Write operations. The only write on the site is the human contact form, protected by Turnstile.