For sixteen years I have given the research away. Every blog post, every guide, every API definition, every scrap of what I have figured out about this space–free, on the open web, no gate. I am not going to stop doing that. But I am finally doing something I have circled for years and kept talking myself out of: I am putting a price on part of my own APIs and MCP servers. Both apis.io and API Evangelist now run a real API and a real Model Context Protocol server, with tiers, keys, and a bill at the end of the month for the people and the agents that want the deep end. I want to be honest about how I am thinking about it, because I have spent a career watching other people get this exact decision wrong.
The principle I keep coming back to is small enough to fit on a sticky note: the map stays free, and I charge for the depth. Discovery is free. Browsing the catalog, searching it, reading the summaries, pulling the tags and the relationships, getting the shape of the landscape–none of that costs anything, and none of it ever will, because a map you have to pay to look at is not a map, it is a toll booth. What I charge for is the synthesis that sits on top of the map: the full research content, the ranked comparisons, the gap analysis, the account-level intelligence, the composites that only exist when you combine a lot of records and do real work to make them useful. The free layer answers “what is out there.” The paid layer answers “what should I do about it.” That second question is the one people and companies have always been willing to pay to have answered well, and it is the only part I feel right about metering.
On the apis.io side, that map is the supply side of the API economy–thousands of providers and their APIs, rated, tagged, searchable, with every machine-readable artifact I can get my hands on. Anyone can hit the API or point an agent at the MCP server and browse all of it for free. What sits behind a key is the synthesis: comparing providers head to head, finding the gaps in an API’s surface, assembling a recommended stack for a set of capabilities, and now the demand-side Insights–what the Fortune 1000 are actually building, buying, and hiring for, and which providers serve that demand. The supply side has always been free to look at. The moment you want it to make a decision for you, that is the moment it becomes worth paying for, and that is the line I drew.
API Evangelist itself is the other half. Sixteen years of writing, guidance, governance building blocks, conversations, and papers are now a queryable API and an MCP server, so a person or an agent can pull the research the same way they would pull any other resource. Discovery is free there too–search everything, read the summaries, follow the related-content graph, get the map of what I have covered. The depth is Pro: the full content bodies, the curated topic guides that bundle a subject end to end, the catalog intelligence across the whole network. I am not charging you to find the research. I am charging you to consume all of it, in bulk, in a form a machine can act on–which is a very different act than a human reading one post over coffee.
The reason the API and the MCP server are the surface I am monetizing, rather than a subscription bolted onto a website, is agents. The consumer has changed. For twenty years I optimized everything for a human developer with context and patience. Increasingly the thing on the other end is an agent doing the discovery and the synthesis itself, programmatically, at a scale no human ever read at. Metered API and MCP access is the honest way to charge for that–you pay for what you pull. An agent that hammers the deep endpoints a hundred thousand times a day is doing real work against infrastructure I have to keep standing up, and asking it to carry some of that cost is not greed, it is arithmetic. The free tier is generous enough to explore and build against; the paid tiers are where sustained, high-volume, decision-grade consumption lives.
I am doing all of this with my eyes open about how it usually goes, because I just wrote about the pattern where everything is open at launch and walled off by Series D. I am trying to run that story backwards. The whole point of putting a price on the depth is to keep the map genuinely, durably free–the paid layer is what funds the open layer, not what replaces it. I have no investors to satisfy and no exit to engineer, which is the entire reason I can make that promise and mean it. The free tier is not a trial that quietly shrinks every quarter until it is useless. It is the product. The paid tier exists so the free one can stay that way.
The newest piece, the demand-side Insights, is where I think the economics actually close. The startups I profile want to know which Fortune 100 accounts look like they need what those startups sell, and why, and when. Media partners want the quarterly story of how enterprise adoption is moving. Those are people and organizations with real budgets and a real reason to pay, and every dollar they spend on the demand-side intelligence subsidizes the open supply-side catalog that everyone else uses for free. That is the shape of the whole thing in one sentence: charge the few who get outsized leverage from the depth, so the many can keep the map for nothing.
None of this is about getting rich, and anyone who has watched me work knows that was never the plan. It is about independence and sustainability–funding sixteen more years of the open, unglamorous, load-bearing work by charging the people and the agents who extract real leverage from the deep end of it. The map stays free. The depth has a price. And if I do it right, you will barely notice the wall is there, because it is standing exactly where the value crosses over from “knowing what exists” to “knowing what to do”–and it is holding the free half up.
