What is APIs.io?
APIs.io is the search engine for the world’s APIs — an open, agent-friendly discovery network that helps people and machines find the right API for the job. Think of it as a public index of the API economy: for every provider on the network, APIs.io gathers up everything that describes how their APIs work — the APIs themselves, their specifications, data schemas, ready-to-run collections, workflows, governance rules, security posture, pricing and rate-limit profiles, and their AI agent surfaces — and organizes it into a single, consistent, searchable catalog. Every provider is then rated on one public rubric, the Kin Score, so the catalog says not only what a provider publishes but how well.
What makes APIs.io different is that it is built to be read by machines as easily as by humans. Every provider is described using open specifications, starting with APIs.json as the root, so the entire catalog is structured, linkable, and machine-readable. That means it works equally well whether you are a developer browsing for a payments API, a search engine indexing the space, or an AI agent grounding itself in real, up-to-date API knowledge. Everything on the network is open-access and explicitly licensed for search indexing, AI grounding, RAG, and model training.
Today the network indexes 31,200+ API providers, 154,000+ APIs, 143,000+ JSON Schemas, 121,200+ API collections, and millions of individual data points across more than forty sections — artifacts, signals, taxonomies and geographies — all kept current from a Git-versioned catalog and rebuilt nightly.
The Specification Ecosystem
Every artifact in the network is described by one or more open specifications. APIs.json is the root — the meta-spec that ties everything else together — and every other spec is declared as a typed property inside a provider’s apis.yml file.
Live counts from the current build — APIs.json at the root, declaring every artifact type across the network. Rebuilt nightly.
| Category | Specification | Role | Site |
|---|---|---|---|
| Discovery | APIs.json | Root meta-specification. Every provider is an apis.yml file that declares what specs it publishes and links to all of them. |
apis.io |
| Description | OpenAPI | REST/HTTP API descriptions — the largest catalog on the network | apis.io/apis/ |
| AsyncAPI | Event-driven, pub-sub, webhook, and streaming API descriptions | apis.io/asyncapis/ | |
| GraphQL | GraphQL endpoint descriptions, documentation, and reference links | apis.io/graphqls/ | |
| Request / Response Examples | Real example payloads lifted from contracts, one page per operation, so a caller sees a correct call before making one | apis.io/examples/ | |
| OpenAPI Overlays | Changes applied on top of an OpenAPI document — phrasing, servers, descriptions, security — each page saying who wrote it | apis.io/overlays/ | |
| OpenAPI Extensions | Every x- extension found across the catalog's contracts — who uses it, for what |
apis.io/extensions/ | |
| HTTP Headers | Request and response headers declared across contracts, by category | apis.io/headers/ | |
| Workflow | Arazzo | Multi-step API workflow specs — chains OpenAPI/AsyncAPI operations into sequences | apis.io/arazzos/ |
| Event Channels | Webhook, streaming and pub-sub channels read out of AsyncAPI documents, by provider | apis.io/channels/ | |
| Schema | JSON Schema | Data model schemas extracted from OpenAPI and AsyncAPI specs | apis.io/schemas/ |
| JSON-LD | Semantic context files that map API concepts to linked-data vocabularies | apis.io/jsonld/ | |
| JSON Structure | Data shapes read out of every contract's schemas — the objects an API actually exchanges | apis.io/json-structure/ | |
| Vocabularies | Each provider's resources, actions, personas and domains as a searchable vocabulary | apis.io/vocabularies/ | |
| Governance | Spectral Rules | API linting rulesets for validating OpenAPI and AsyncAPI descriptions | apis.io/rules/ |
| Standards | Interfaces that became standards by being copied — S3, the OpenAI and Anthropic APIs, CKAN, Ethereum JSON-RPC — each profiled as a full artifact set | apis.io/standards/ | |
| Security | Security Posture | Vulnerability disclosure, domain security (DNSSEC, SPF, DMARC, HSTS) and trust-center profiles, probed rather than claimed | apis.io/security/ |
| OAuth Scopes | Every scope a provider declares, with its flows and descriptions, so a least-privilege grant can be chosen | apis.io/scopes/ | |
| Agentic Access | How an agent gets in: the machine onboarding path, credential shapes and gates, apart from the human one | apis.io/agentic-access/ | |
| Collections | Open Collection | Open, vendor-neutral collection format (.opencollection.json) |
apis.io/collections/ |
| Postman Collection | Ready-to-run request collections in Postman format | apis.io/collections/ | |
| Operational | Plans | API pricing plan profiles (API Commons Plans format) | apis.io/plans/ |
| Rate Limits | Rate-limit policy profiles (API Commons Rate Limits format) | apis.io/rate-limits/ | |
| FinOps | Cost and billing profiles aligned with the FinOps Foundation FOCUS framework | apis.io/finops/ | |
| Agent | Agent Skills | SKILL.md files published by providers — task-shaped instructions ready for direct agent use (APIs.io publishes twenty of its own) |
apis.io/agent-skills/ |
| MCP (Model Context Protocol) | MCP server declarations — tools, auth, and connection details — probed for a server that answers, and marked when the catalog derived the card rather than the provider | apis.io/servers/ | |
| A2A Agent Cards | Agent-to-agent cards served at /.well-known/agent-card.json, graded on conformance rather than presence |
apis.io/a2a/ | |
| Agentic Resource Discovery | The /.well-known/ard.json manifest naming every agent-facing surface a provider runs |
apis.io/ard/ | |
| Agent Registry | Agents that registered their own card with the network, with proof they control the domain that serves it | apis.io/agents/ |
How It Works
A central build pipeline reads from a set of provider repos at github.com/api-evangelist — Git-versioned APIs.json profiles, one per provider — and emits a Jekyll collection for every section. Each section builds independently and the whole network is served from one origin at the apis.io apex (apis.io/providers/, apis.io/apis/, …), rebuilt nightly with a full rebuild and rescore every Friday.
Every provider and API page carries:
- A generated prose overview derived deterministically from structured frontmatter
- Schema.org JSON-LD (
Organization,WebAPI,Dataset,CollectionPage) - Open Graph and Twitter Card metadata
- A Kin Score — a composite 0–100 rating banded Exemplar / Strong / Developing / Thin / Emerging / Minimal — and a standalone Agent Readiness rating (Agent-Native / Agent-Ready / Agent-Aware / Human-Only), both recomputed from a public rubric on every build, with every check it reads named
- Cross-page internal links to related providers, industries, regions, and specs
The network also publishes machine-readable feeds at the apex: llms.txt, sitemap.xml, an RFC 9727 api-catalog linkset, an MCP server card, an A2A agent card and an ARD manifest. Start at apis.io/.well-known/api-catalog or apis.io/llms.txt; the blueprint lists every file we serve and what each is worth.
Beyond the catalog
- Browse by market, not only by provider. Tags, industries, areas of API practice, topics, business capabilities, use cases, regions and countries, regulations, venture-capital portfolios, the secondary market, Gartner Magic Quadrants and corporate API estates — each a cohort of providers scored on the same rubric, with an auto-generated brief that states its own coverage.
- The API and MCP server. Everything on this site is available as JSON at
apis.io/api/v1and as 130+ tools onapis.io/mcp. The catalog, any single provider, tags and artifacts are free and need no key; synthesis across the catalog — ratings, cohorts, capabilities, gap analysis, stack design — needs a key on the Understanding or Influence tier. Developer docs · how to get access. - For an agent that is new here. The playground is a short list of APIs that need no key and cost nothing, re-proved nightly. KYA is how the network knows who is calling; the trust layer is how an agent gets in and proves who it is.
- For a provider. Add your API, read your own rating and what would move it, copy the agent surface we run ourselves, and register your agent card. Corrections are free and always will be.
- Insights — what companies build, buy and hire for, read from the catalog. Your API stack — watch the providers you depend on for retirements and changes.
Adding an API
Anyone can add an API at apis.io/add/. Give us the name, your website or APIs.json URL, where the documentation lives and an email we can reply to, and a person reads it. Listings are published as Git-versioned APIs.json profiles under github.com/api-evangelist and flow into the catalog on the next build. If the catalog is wrong about you, say so — corrections and disputes are free and are read by a person.
Open Access
The APIs.io catalog is licensed for public use including search indexing, AI grounding and RAG, and model training. See apis.io/terms/ for the full terms.
Team & Support
APIs.io is a project led by Kin Lane (API Evangelist) and Steve Willmott (Timewarp Labs). It is built on APIs.json and API Commons, both maintained by Kin and Steve as open infrastructure for the next generation of API discovery.
The upstream provider catalog is maintained at github.com/api-evangelist and the network infrastructure lives at github.com/api-search.
Questions or issues: info@apis.io, the support page, or an issue in the APIs.io Inbox.
