Find and compare AI agents, skills, MCP servers and templates, across marketplaces
Updated
We compare the stores that sell agents, skills, MCP servers and templates, link straight into each store's search, and trust-score 1,576 MCP servers and 661 x402 APIs daily. Free.
New AI agent tools, trust-scored, free
Plus one rising trend and one agent build you can copy.
We check 2,244 agent tools daily. See a sample(count as of Oct 24, 2026)
MCP servers, x402 APIs and agent platforms. One search.
Where to find and sell agents, skills, MCP servers and templates, with each store's fees and a link into its own search. We link to each store; we never sell or take payments for their listings.
Search MCP servers and x402 APIs, trust-scored daily Directory →
2,244 agent tools, including 1,576 MCP servers and 661 x402 pay-per-call APIs, each with an automated trust and uptime score. (counts.json, as of ; how scoring works)
Search tweets with advanced operators (recommended). Body param is `query` (not q), e.g. {"query":"from:elonmu
x402 · $0.005
glim.sh
CDartzDabz Crew BuilderOur own Grok Bot template
A CEO bot that asks a few questions about your online business, then builds a crew of up to 8 worker bots in any niche: site, SEO, research, social, money, newsletter, clips, and AI-search tracking. Nothing sends, posts, or spends without your OK.
How we score MCP servers: readiness plus trust and uptime checks, with official vendors pinned first.
A vertical explainer for the DartzDabz best MCP servers page. A card reads 'Overall score 0 to 100', built from readiness checks and trust and uptime checks, and official vendors are pinned first.
Your AI agent just wrote code. Where does it actually run?
Where AI agent code actually runs: microVM sandboxes, a red-team test, and the approval step OWASP recommends.
A control-room board explains AI agent sandboxes: microVM boot time and overhead, a red-team test of sandbox escapes, and OWASP's advice to sandbox agent code and require human approval for high-impact actions. Sourced figures are on the AI agent sandboxes page.
What the free DartzDabz newsletter covers, and how a human reviews each issue before it goes out.
A control-room board shows how the DartzDabz newsletter is put together from new tools, trust-score changes, a rising trend and a setup to copy, with a human review step before sending.
Cursor mcp.json for GitHub's official MCP server: point it at api.githubcopilot.com/mcp/ and add your GitHub token as an Authorization header. dartzdabz.com/best-mcp-serve…
Mastercard just picked x402 for agent payments. So what does an x402 call actually cost? We track 623 x402 endpoints, 574 scored daily. Median price across scored endpoints today: $0.0072 a call. Grades, latency and price for each one: dartzdabz.com/x402-apis/?utm…
Which MCP servers actually answer? We check remote MCP servers every day and rank them on readiness, trust and uptime. Here's how the score works: dartzdabz.com/best-mcp-serve…
x402 prices: "median across scored endpoints" uses endpoints with a current score and a listed price; the monthly report uses all priced endpoints. How we calculate it.
Guides
Plain answers on MCP, x402, Agent Skills and the tools agents use, with sources.