Docs
API
One remote MCP server. Your AI calls the tools below; the same calls exist as HTTP JSON under /v1.
Connect
claude mcp add --transport http mhtirogla https://mhtirogla.ai/mcp
Or add https://mhtirogla.ai/mcp in any MCP client. No key needed to read; an API key from your account raises the limits.
Tools
| MCP tool | HTTP | What it returns |
|---|---|---|
search | GET /v1/search?q= | Systems and building blocks for a query, each with why it matched |
system_graph | GET /v1/system/{id} | The component graph of a system: every part, what it builds on, alternatives and design questions |
get_entry | GET /v1/entry/{slug}?part= | A page: overview, or one section (Code, Pitfalls, Efficiency, …), with its relations |
list_systems | GET /v1/systems | All systems by section |
list | — | Block pages by domain |
request_entry | POST /v1/requests | Ask for a missing topic; repeats count as votes |
Example call
search{query: "rate limiter for an API"} returns
system global-rate-limiter
system api-gateway
block distributed-rate-limiting Enforce one rate limit across N servers: a central atomic counter,
a limit split per node, or batched token leases…
and get_entry{slug: "distributed-rate-limiting", part: "overview"} returns the page's overview.