Index / Verified tools

Verified MCP tools, by capability

Every other registry indexes servers: a name, a URL and whatever prose the publisher wrote. The tool names and input schemas an agent actually has to match on exist in no index. So we ask each server directly, and store what it answers.

Analytics and monitoring
8,072 tools · 3,616 servers
Maps, location and weather
7,951 tools · 3,319 servers
Files and storage
6,977 tools · 3,630 servers
Calendar and scheduling
5,871 tools · 1,572 servers
Science and research
5,435 tools · 2,697 servers
Payments and billing
5,337 tools · 2,606 servers
GitHub and version control
4,975 tools · 2,576 servers
Crypto and blockchain
4,842 tools · 1,383 servers
Web scraping and browser automation
4,294 tools · 1,953 servers
Legal and government data
4,209 tools · 740 servers
Images, audio and video
2,676 tools · 556 servers
AI models and embeddings
2,075 tools · 832 servers
HR and recruiting
1,943 tools · 533 servers
Project management
1,930 tools · 1,736 servers
E-commerce
1,820 tools · 505 servers
PDF and document extraction
1,763 tools · 532 servers
Cloud and DevOps
1,555 tools · 271 servers
Finance and market data
1,430 tools · 503 servers
Email
1,207 tools · 426 servers
Social media
1,207 tools · 162 servers
Databases and SQL
993 tools · 214 servers
CRM and sales
634 tools · 268 servers
Security and compliance
570 tools · 219 servers
Messaging and chat
454 tools · 144 servers
Web search
430 tools · 157 servers
Translation and language
191 tools · 80 servers

What is in the index

Measured, not claimed. This is the whole index by kind, 111,579 resources in total.

MCP servers103,667
Agent skills3,894
OpenAPI services2,830
Other callable resources492
Machine-readable docs429
A2A agents119
Plugin manifests105
Catalogues18
ARD registries14
Packages6
GraphQL APIs2
Agent descriptors (ACP, OASF, AgentFacts)1
Datasets1
WebMCP browser tools1

How the ecosystem spells the same thing

Media types as publishers actually wrote them. Filters match exactly, so a registry that does not normalise these silently drops thousands of entries. We normalise all of them, plus both URN prefixes in circulation (urn:air: and urn:ai:), on ingest.

application/mcp-server-card+json99,579
application/mcp-server3,570
application/ai-skill2,602
application/vnd.oai.openapi+json2,596
application/ai-skill+md1,209
application/mcp-server+json512
application/json397
application/openapi+json201
text/markdown166
text/html118
application/a2a-agent-card+json103
application/vnd.anthropic.claude-plugin+json82
This is not a directory of every server. We measured what ranks for MCP queries and the primary source wins: mirroring it at scale adds nothing. These pages exist because the verified tool surface is data that is not published anywhere else.

The whole corpus is also an open dataset under CC BY 4.0: huggingface.co/datasets/AgenticResourceDiscovery/verified-mcp-tools.