How AI is applied across API Evangelist and APIs.io. Read my AI disclosure →
API Evangelist API Evangelist
Discovery
Learnings
Guidance
Toolbox
Alignment
API Evangelist LLC

Search Sixteen Years Of API Research From One Call

October 30th, 2026 · Kin Lane
Search Sixteen Years Of API Research From One Call

This is the second story in my series on what each API Evangelist operation can do. Yesterday I linted an OpenAPI. Today I’m going the other way, from checking a contract to finding out what’s already known.

Say you’ve been asked to write your company’s API deprecation policy. You could start from a blank page, or you could see what has already been written, what rules can enforce it, and who has done it well. The MCP tool for that is search_api_evangelist, and the REST operation behind it is GET /search. It searches the whole API Evangelist network at once, and it’s free.

What I gave it

One word, deprecation, first across everything and then narrowed by content type:

{ "q": "deprecation" }
{ "q": "deprecation", "types": ["rules", "policies", "guidance"] }
{ "q": "deprecation", "types": ["posts"] }

What came back

The search covers 17,140 records. One word narrowed that to 103, and the response counts them by type, so before reading a single result you can see the shape of what exists:

posts       68     guidance     5     rules        2
standards    7     areas        3     lifecycle    2
policies     6     vocabulary   4     properties   2
strategies   2     tools        1     experiences  1

The top of the list was the concept itself: the Deprecation lifecycle stage and the Deprecation topic area. Filtering to governance content returned 13 records, and that’s where this gets useful for my policy writer:

The posts filter returned 68 stories, ranging from good examples of deprecation policies in 2015, to EVRYTHNG’s deprecation page in 2018, to using OpenAPI Overlays for deprecation and migration this August.

Search isn’t perfect, and you should know where it gets weaker. When I searched tools, solutions, standards, papers, conversations and videos together, I got eight results. A few were clearly relevant, like Versioning Protocols and CHANGELOG.md. Others, like Mobile Connect and Product Backlog, were much looser matches. It also found no paper, conversation or video on this word. Narrow by type, and read the results rather than trusting the count. Those zeros only mean this word didn’t match, so try sunset or versioning too.

What that makes possible

  • Start a policy from what exists. Going from one word to guidance, policies and an enforceable rule takes three calls, which is a better first draft than a blank page.
  • Size up a topic first. The type counts tell you whether a subject is mostly writing, or whether it has also become policy and rules that can be enforced.
  • Give an agent a grounded source. An agent writing governance documentation can search first, then fetch a full record and link to it, rather than relying on what it remembers.
  • Walk from idea to enforcement. Use types to move from stories, to guidance, to policies, to the rules you can run in CI.

Try it

It’s free and needs no key. From any MCP client, connect to https://mcp.apievangelist.com/mcp (or the lighter https://mcp.apievangelist.com/mcp/research) and call search_api_evangelist. Over REST:

curl 'https://api.apievangelist.com/v1/search?q=deprecation&types=rules,policies,guidance'

The operation is documented in the API reference, and the MCP server docs explain how to connect a client. Pick the thing you’ve been asked to write next, search for it, and see how much of it you don’t have to start from scratch.