Remote desktop control
Live VP8 screen capture streamed peer-to-peer, with mouse and keyboard injected on the host. The viewer can watch read-only or take full control.
Live
v1.3.0 running at remote.etherneco.com
Remote desktop and support,
Remote desktop and support,
on infrastructure you own
etherRemote AI connects a technician's browser directly to a remote Windows machine over peer-to-peer WebRTC — with passkey sign-in, granular session permissions, an audit trail and a built-in AI copilot. No third-party cloud in the media path.
Self-hosted
Peer-to-peer media
No agent install for the viewer
Local-AI ready
- WebRTC
- VP8 screen stream, encrypted with DTLS-SRTP
- 3 ways in
- Password, TOTP 2FA, or passkey single sign-on
- 1 executable
- Self-contained host agent — no runtime to install
- 100% yours
- Runs on your Windows Server or Linux hosts
What etherRemote is
One platform for remote access, remote support and screen sharing
etherRemote AI is a remote desktop and remote-support platform you deploy yourself. A small agent runs on each machine you want to reach; technicians connect from an ordinary web browser. The server brokers the introduction, then the screen stream and the keyboard and mouse traffic flow directly between the two endpoints.
That architecture is the point. Because media is peer-to-peer and the whole stack runs on your own hardware, the pixels of a finance workstation or a hospital terminal never traverse a vendor's cloud — which is what makes it viable in environments where conventional remote-support tools are simply not allowed.
On top of the connection sit the things support teams actually need day to day: a machine inventory with live presence, temporary share links for one-off help, file transfer, clipboard sync, chat, role-based permissions, an append-only audit log — and an AI copilot that can run entirely on a local model.
- For internal IT — reach every managed workstation and server from one console.
- For MSPs and helpdesks — send a Quick Share link; the customer joins in a browser with no account and no installer.
- For regulated environments — keep the data plane, the identity provider and the audit trail inside your own perimeter.
At a glance
- Controlled machine Windows 10/11 · Server
- Technician side Any modern browser
- Screen codec VP8 over WebRTC
- Signalling SignalR / WebSocket
- NAT traversal STUN + self-hosted TURN
- Identity JWT · TOTP · OIDC passkeys
- Database SQLite · MySQL · PostgreSQL
- Server runtime .NET 9 / ASP.NET Core
- Deployment Windows Server or Docker
- Licence Proprietary
Capabilities
Everything a support session needs
Built as separate modules on a shared core, so you can run the whole platform or just the parts you need.
Quick Share links
Generate a temporary, token-scoped URL for a single session. The guest opens it in a browser — no account, no download, and the link expires on its own.
AI copilot
Summarise a session, explain an error and draft the closing support note. Runs against a local Ollama model or any OpenAI-compatible endpoint — your choice.
File transfer
Move files to and from the remote machine during a session, with object-storage backing for larger payloads and an administrative view of every exchange.
Chat & clipboard sync
Talk to the person at the other end without leaving the session, and copy text between the two machines as if they shared a clipboard.
Machine inventory
Every enrolled machine reports its real operating-system caption and live online/offline presence, so you know what you are connecting to before you connect.
Access control & audit
Organisations, roles and per-session permission flags decide who may view, control, transfer files or elevate — and every action lands in an append-only log.
VR workspace
A WebXR client puts the remote screen on a virtual display you can reach from a headset — an independent app riding the same signalling gateway. Preview
Passkey sign-in
Sign in with a fingerprint, face or device PIN through EtherLogin single sign-on — or keep email and password with TOTP two-factor. Both are live today.
The product
Screens from the running system
Captured from the live deployment at remote.etherneco.com — not mock-ups.
Both captures are of a real deployment — our own test fleet, not mock-ups. Windows, macOS
and Linux hosts sit side by side in the same list, and the session above is a Mac driven from a
browser. The machine names and every desktop you can see — the thumbnails and the streamed
session alike — are blurred because the hosts are real ones; nothing else is retouched.
Book a walkthrough and we will drive a live session for you.
How it works
The server introduces; the peers do the work
Signalling is centralised so sessions are controlled and logged. Media is not, so it stays fast and private.
1
Enrol the machine
The host agent runs on the Windows machine you want to reach, authenticates with a machine key and registers itself. An administrator decides whether it is allowed to accept sessions.
2
Start a session
A technician picks the machine from the console, or sends a Quick Share link to someone who has no account at all. The gateway pairs the two sides over a WebSocket.
3
Negotiate a direct path
The peers exchange WebRTC offers, answers and ICE candidates through the gateway, then connect directly. When a firewall forbids that, a TURN server you host relays the traffic instead.
4
Work — and record what happened
Screen frames flow one way; input, clipboard, chat and file events flow back over a data channel. Permissions are enforced per session and every action is written to the audit log.
Supported systems
What runs where
Two different questions: which machines can be controlled, and which devices can do the controlling.
| Role | Operating systems | Requirements | Status |
|---|---|---|---|
| Controlled machine host agent |
Windows 11 · Windows 10 (x64) Windows Server 2016 · 2019 · 2022 · 2025 | Self-contained executable — no .NET runtime to install. Screen capture uses GDI; input uses the native SendInput API. |
Supported |
| Controlled machine host agent |
macOS · Linux | The agent already builds and runs cross-platform; the capture and input back-ends for these systems are not implemented yet. | On the roadmap |
| Technician side web console |
Windows · macOS · Linux · ChromeOS iPadOS · iOS · Android | Any browser with WebRTC and WebSocket: Chrome and Edge 111+, Firefox 115+, Safari 16.4+. Nothing to install. | Supported |
| Technician side desktop client |
Windows (shipped) · macOS and Linux buildable | A native client built on Avalonia for teams that prefer an application window over a browser tab. | Windows first |
| Server self-hosted |
Windows Server 2019+ behind IIS Linux, natively or via Docker Compose | ASP.NET Core on .NET 9. Ships with SQLite for small deployments; MySQL or PostgreSQL for larger ones. | Supported |
| Headset VR workspace |
WebXR-capable headsets — Meta Quest and similar | A separate WebXR application that fronts the same signalling gateway for live remote control in VR. | Preview |
To be precise about today's limits: the machine being controlled must run Windows.
The machine doing the controlling can be anything with a current browser — that half is
genuinely cross-platform.
Security
Designed for networks that say no
Remote-access software is a standing invitation to attackers. These are the controls that are implemented and running today.
Encrypted media
WebRTC negotiates DTLS-SRTP, so the screen stream and input channel are encrypted between the endpoints — not merely to the server.
Passkeys & 2FA
Passkey single sign-on over OIDC, or email and password hardened with TOTP two-factor. Signed JWTs carry the session.
Closed registration
Accounts exist only by invite code. Machines are enrolled with a key and must be approved before they can accept a session.
Granular permissions
View, control, clipboard, file transfer and elevation are separate flags. The host can revoke control or end the session at any moment.
Append-only audit
Who connected, to what, from where and when — written once and never mutated, so the record survives the incident.
Abuse limits
Per-IP rate limiting across the whole API, tightened sharply on sign-in and the public contact form to blunt brute force and spam.
Hardened transport
TLS everywhere, HSTS, forced HTTPS, and strict response headers — no sniffing, no framing, no referrer leakage.
Patched dependencies
The media and data stacks are tracked against published advisories and updated deliberately, not incidentally.
Honest about maturity. etherRemote is an actively developed product, and parts of the
security model — application-level payload encryption and long-term audit storage among
them — are still being hardened toward production grade. We are happy to walk a security
reviewer through exactly what is implemented and what is not.
Under the hood
Built on boring, durable technology
No exotic dependencies, no proprietary protocol on the wire — WebRTC, WebSocket and HTTP.
Server
- Runtime .NET 9 / ASP.NET Core
- Real-time SignalR
- Media stack SIPSorcery WebRTC
- Data access Entity Framework Core
- Object storage S3-compatible
- Cache / backplane Redis
Clients
- Web console React 18 + TypeScript
- Build tooling Vite
- Desktop client Avalonia 11
- Host agent .NET, self-contained
- Screen codec VP8
- VR client WebXR
Platform
- Identity JWT · TOTP · OIDC
- NAT traversal STUN + TURN
- Databases SQLite · MySQL · PgSQL
- Local AI Ollama
- Hosted AI OpenAI-compatible
- Containers Docker Compose
Roadmap
Shipped in phases, in the open
Each phase landed as a working, tested increment rather than a promise.
✓ Foundation
✓ Session signalling
✓ Screen capture
✓ Remote input
✓ File transfer
✓ Chat & collaboration
✓ AI copilot
✓ Enterprise admin
✓ Quick Share
✓ Passkey sign-on
◷ Security hardening
◷ VR/AR workspace
macOS & Linux hosts
Contact
Let's talk about your deployment
Tell us how many machines you need to reach and what your security team requires. We will come back with a straight answer about whether etherRemote fits — including when it does not.
-
Email us directly
-
A live walkthrough
Roughly 30 minutes, driving a real session against a real machine — no slideware.
-
Security review welcome
Bring your questionnaire. We would rather answer it now than at procurement.
See it running against a real machine
Thirty minutes, your questions, an actual remote session.
Request a demo