Discover what's happening at the edge.

EdgeDisco discovers AI applications, running agents, local model runtimes, agent frameworks, and MCP servers across endpoints, without collecting the conversations inside them.

Scan your edge How discovery works

12,480 endpoints scanned312 signals found0 conversations stored

AI app

Agent

MCP

Optical Engine StatusREADY [SCAN_MODE_ACTIVE]

Live discovery

What's running out there?

scanning edge network

Agent active

orchestrator-agent

edge-node-04 : 4833

Multi-step task runner coordinating tool calls and local memory.

runtime: python 3.12signal #a1f2

Model active

local-llm-runtime

edge-node-11 : 8080

On-device inference serving quantized models with zero cloud egress.

runtime: llama.cppsignal #c7d0

MCP Server active

mcp-file-tools

edge-node-02 : 9000

Model Context Protocol server exposing file and search capabilities.

runtime: node 20signal #b3e9

Coverage

See the whole edge at a glance.

AI Apps

detection

Agents

live status

Runtimes

local models

MCP

servers

Privacy by design

See the shape, not the words.

EdgeDisco maps metadata, protocols, ports, and runtime fingerprints. Prompt and response content stays where it belongs: on your endpoints.

conversation content // excluded

Open source

Inspect every reflection.

Audit what EdgeDisco sees, extend the discovery taxonomy, and run the scanner across your own fleet.

View on GitHub

Discover the edge

Ready to see what's running?

Start discovering