Calm command surface
Ask Hermes, launch workflows, inspect artifacts, or cancel risky work from a focused dashboard — not a wall of JSON.
Open gateway · Local-first · Self-hosted Hermes
An AI command center for your own Hermes agents.
Vylen pairs a tiny on-device connector with a calm, mobile-first control surface. Run Hermes agents on your own hardware, ask them to do work, monitor long-running runs, and stay in command — without handing over your data or your keys.
Product
Your agents. Your hardware. Your call.
Vylen Agent is the controller for self-hosted Hermes Agent instances. It gives you a single, polished surface to see what your agents are doing, decide what comes next, and resume runs from any device — without giving up local ownership.
Outbound-only tunnel
The connector opens a single WebSocket out. No inbound ports, no exposed services, no static IPs required.
Multi-instance ready
Home, work, server, experiment — pair any number of Hermes instances and switch between them with a compact instance chip.
Private by design
Vylen Cloud never stores your Hermes API key. The sidecar injects it locally; cloud only brokers tunnels and routes commands.
Runs, not chats
First-class long-running runs with progress, status, and resumability — purpose-built for agent workloads, not chat transcripts.
Open gateway
The Hermes-side gateway plugin is open source MIT. Read it, fork it, audit it, run it yourself — no vendor lock-in on the part that runs next to your data.
How it works
Three pieces. One outbound tunnel.
Run a connector
Drop the hermes-connect sidecar next to any Hermes instance. It pairs with a one-time token, then opens a single outbound WebSocket to Vylen Cloud.
Pair from the app
Open the Vylen web or mobile app, sign in, paste the pairing token, and the instance appears as connected — with live health and heartbeat.
Ask it to do work
Launch runs, watch tool activity stream, and resume from any device. Cloud routes commands through the tunnel; your keys never leave the connector.
Latest
From the Vylen blog
Same chat, every screen
Earlier this week I had the same conversation open on three screens at once. Linux Chrome at my desk. Windows Chrome on a second machine. The Vylen iOS app on my iPhone, propped …
The Scary Part: Hermes Approval Depends on Which Tool the Model Picks
While testing Vylen’s new MessageEvent chat path, we hit a behavior that is easy to miss and hard to unsee: In Hermes, whether a dangerous operation asks for approval can …
Hello, Vylen
Most AI tools today ask you to upload your context, your keys, and your work to someone else’s machine. Vylen takes a different bet: your agents should run on your hardware, …