← All docs

Getting started

Getting started

You do not need to be technical to use this explorer. Think of it as a map of the machine-payable web — paid APIs and tools that software agents (and the products behind them) can discover and use.

What you are looking at

The home page shows a snapshot of the market: how many capabilities we track, how they group into niches, and where real money has moved recently. The numbers update as we index more of the ecosystem.

Your first five minutes

  1. Start on the home page — read the headline figures, then pick Explore the catalog or open Search in the top navigation.
  2. Search for something you care about — try a plain phrase like weather, image generation, or payments. Results are ranked by relevance. Use the filters on the left if the list feels too broad.
  3. Click any result — you will see what the capability does, what it costs, and whether we have successfully reached it recently.
  4. Browse Niches — related capabilities grouped into markets (for example, all weather APIs in one place). Each niche shows how much demand we can attribute to it.
  5. Check Opportunities — niches where demand looks high but few providers exist. Useful if you are deciding what to build next.

The main sections, in plain English

Section What it is for
Search “Find me a capability that does X.”
Niches “Show me the whole market for X.”
Opportunities “Where is demand ahead of supply?”
Wallets “Follow the money — who is actually getting paid?”

Reading the money figures

You will see two kinds of dollar amounts on niche and wallet pages:

  • Proven demand — volume we can tie to specific capabilities with high confidence.
  • Addressable opportunity — an upper bound on what could have flowed through a niche when payments are ambiguous (many capabilities share the same price on the same wallet).

Do not add “addressable opportunity” numbers across niches and treat the total as real revenue. They are guides for comparison, not audited books.

Tips if you feel lost

  • Too many results? Add a filter (source, category, price range) or type a more specific search.
  • Empty niche or wallet page? The pipeline may still be indexing — check back after the home-page figures populate.
  • Building software or agents? See API, MCP, and llms.txt in the docs for how to connect programmatically.

Who this is for

  • Product and strategy — see what exists, what earns, and where gaps are.
  • Investors and analysts — market sizing with honest bounds, not false precision.
  • Builders and agents — find and integrate paid capabilities without maintaining your own catalog scrape.

Classification confidence

Each capability gets an automatic semantic tag (domain and facet). Tags marked Low confidence or Review were uncertain enough that a second model pass may have been used; Human assigned means a curator overrode the model. Curators set overrides directly on the endpoints table (human_semantic_facet, human_assigned_at, human_assigned_by); those overrides persist across re-scrapes until changed.