Hashnet
Search and compare agents across MCP, NANDA, A2A, and OpenRouter registries before you commit to an architecture or vendor.
Overview
Hashnet is an Idea-phase MCP server that searches 59,000+ AI agents across MCP, NANDA, A2A, and OpenRouter for discovery and comparison.
What is this MCP server?
- Search 59,000+ agents across MCP, NANDA, A2A, and OpenRouter from one MCP server
- npm package @hol-org/hashnet-mcp (v1.0.24) with stdio transport
- Registry Broker API key from hol.org for authenticated discovery
- Cross-protocol lookup so you are not siloed to a single agent marketplace
- Catalog scale stated as 59,000+ agents
- Protocols: MCP, NANDA, A2A, OpenRouter
- npm @hol-org/hashnet-mcp version 1.0.24, stdio transport
Community signal: 12 GitHub stars.
What problem does it solve?
Agent registries are split across protocols and marketplaces, so picking the right remote agent for your product means manual tab-hopping and stale bookmarks.
Who is it for?
Indie builders planning multi-agent or MCP-heavy products who need a single discovery surface before Build integrations.
Skip if: Teams that already standardized on one in-house agent with no external registry needs.
What do I get? / Deliverables
With Hashnet MCP configured, your agent can query a unified discovery index across major agent ecosystems and narrow options before you integrate.
- Unified agent search results across MCP, NANDA, A2A, and OpenRouter
- Shortlists and metadata your agent can use in architecture notes or integration plans
- Repeatable discovery queries without manual registry websites
Recommended MCP Servers
Journey fit
Choosing which agents and protocols to build on happens in Idea when you are still exploring the ecosystem, not after infra is frozen. Hashnet is universal agent discovery—Discover is the shelf for tooling that helps you find existing agents and registries to reuse or integrate.
How it compares
Cross-registry agent search MCP, not a skill that implements business logic inside your repo.
Common Questions / FAQ
Who is Hashnet MCP for?
It is for solo and small-team builders evaluating third-party AI agents across MCP, NANDA, A2A, and OpenRouter before they commit to an integration path.
When should I use Hashnet MCP?
Use it during idea and discovery when you need to find, compare, or shortlist external agents instead of building every capability from scratch.
How do I add Hashnet MCP to my agent?
Add the stdio server for @hol-org/hashnet-mcp (npm), set REGISTRY_BROKER_API_KEY from hol.org if required, and restart your MCP client.