TunnelMind — observability for the agentic internet

The web is filling up with traffic no human ever typed. TunnelMind watches the hostile slice of it — a distributed fleet of attested sensors plus public feeds, every observation Ed25519-signed at the source so the data can prove where it came from.

One signed corpus, three lenses

Scry — signed observations of hostile network actors. IPs, ASNs, behaviors, threat-feed overlap, actor classification. api.tunnelmind.ai
Sigil — programmatic-advertising supply-graph trust layer. Publishers, SSPs, DSPs, seats, OpenRTB SupplyChain, entity trust scoring. data.tunnelmind.ai
Tracker — the demand-side graph of who is watching whom on the open web; recurring SDK fingerprints and tracker-operator entities.

The moat is the join: a fused cross-lens verdict on any entity (attack signal × trust signal × tracker signal) no siloed incumbent can compute.

For autonomous agents

MCP servers (Streamable HTTP, JSON-RPC 2.0, all listed in the official MCP registry):

mcp-data.tunnelmind.ai — ai.tunnelmind/data (43 tools)
mcp.tunnelmind.ai — ai.tunnelmind/scry (12 tools)
mcp.sigil.tunnelmind.ai — ai.tunnelmind/sigil (11 tools)

Discovery: ai-services.json, agent-manifest.json, llms.txt.

Standards

This page is the JavaScript-free fallback. The full live radar runs in a browser with JavaScript enabled.