Remote access for developers and their AI
Secure remote access to your dev machine.
Install one CLI, paste one URL into Cursor, Claude, or ChatGPT, and your AI works on your machine from anywhere. It only connects outward, so there is no port forwarding and no VPN.
Connect in three steps
1Install on the machine you want to reach
curl -fsSL https://cli.flexx.dev/install.sh | sh2Sign in and bring it online
flexx setup && flexx connectOpens your browser to sign in. The machine shows as online in your dashboard within seconds.
3Paste the MCP URL into Cursor, Claude, or ChatGPT
https://relay.flexx.dev/v1/mcpOAuth asks you to approve access the first time a client connects.
Outbound only
Your machine dials the relay. No port forwarding and no VPN.
MCP or SSH
Any MCP client, or ssh flexx-yourmachine from a terminal.
Free to start
Unlimited remotes and Memory with no credit card.
Three steps. Stop after any of them.
Each step unlocks something new. Memory, for example, works the moment you have an account.
Install the CLI on a machine
One command on the machine you want AI agents to use. It becomes a remote, with no VPN and no port forwarding.
curl -fsSL https://cli.flexx.dev/install.sh | shThe flexx binary is on your machine
Run setup, then connect
This signs you in, registers the machine, and starts the daemon. The machine shows as online in your dashboard within seconds.
flexx setup && flexx connectYour machine is online on the relay
Paste the MCP URL into your AI client
Works in Cursor, Claude Desktop, ChatGPT, or any MCP client. OAuth handles sign-in for you.
https://relay.flexx.dev/v1/mcpYour agent can run tools on your machine
Only need agent Memory? It works with just an account and no install: create one.
Built so you stay in control.
Giving an AI access to a machine is a trust decision. Flexx keeps it narrow, visible, and revocable.
Outbound-only connection
Your machine dials the relay over WebSocket. You never forward a port, open a firewall rule, or run a VPN.
You approve every client
AI tools connect with OAuth 2.1 and PKCE. You choose which machine each client can reach and can revoke access from the dashboard at any time.
Safe by default on the machine
Path scoping, a command blocklist, and file size limits apply out of the box, so an agent stays inside the workspace you gave it.
Every tool call is logged
See what your agents ran, on which machine, and whether it succeeded, in the Activity page of your dashboard.
Remote is the core. Add the rest when you want them.
Router, Memory, and Agent are optional extras. Each is a standalone MCP server, and Flexx Agent wires them together for you.
Remote
Give any AI real hands on your machine
An MCP server for Cursor, Claude, Grok, or any client. Agents run code, edit files, and execute commands on your machine with zero inbound ports.
Install the CLI on the machine you want to reach.
Router
One place for every model
Bring your own models and providers, set budgets, and use one OpenAI-compatible endpoint from anywhere, including inside Flexx Agent.
Add your own provider API keys.
Memory
Agents that remember across tools and sessions
A shared knowledge graph over MCP. Save decisions, preferences, and project context once, then recall them from any client.
Nothing. Any account works, with no install.
Agent
The assistant that uses all three
One assistant that already has hands, a model layer, and durable memory. You chat, and it acts, remembers, and picks the right model.
A model, either yours or your own keys.
Your keys. Your models. One endpoint.
Add a provider key in the dashboard and every model it can reach shows up in one place. Flexx Agent uses them, and your own code can call the router too. Provider keys stay on the server.
- Anthropic
- OpenAI
- OpenRouter
- Groq
- Any OpenAI-compatible endpoint
Use it from your own code
Create a router key in the dashboard, then point any OpenAI or Anthropic client at your router URL.
Chat completions use the OpenAI format and messages use the Anthropic format. Per-key budgets and allowlists are enforced before any request reaches a provider.
What your agent can actually do
MCP protocol
Agents use standard MCP tools for file operations, shell commands, screenshots, and more.
SSH access
Run ssh flexx-myserver like any other host. Proxy settings and known hosts are configured for you.
Localhost tunneling
Opt-in: share a dev server at a public URL with one command. Nothing is shared unless you start a tunnel.
Screenshots and vision
Agents capture web pages to check the UI and run visual tests.
Workspace safety
Path scoping, a command blocklist, and file size limits keep your machine safe by default.
Auto-discovery
flexx finds ripgrep, Docker, Jupyter, Chrome, GPUs, and other tools by itself.
Agent memory
Graph-backed memory for preferences, decisions, and project knowledge, with semantic search over MCP.
Free to start. No credit card.
Cloud Free
$0forever
Hosted relay and one-click OAuth for individual developers.
- Unlimited remotes
- Knowledge graph memory
- Encrypted reverse tunnels
- Audit logs and vision
Team
Custom
For teams that share remote development environments.
- Multi-user organizations
- Custom tunnel domains
- SSO
- Dedicated support
Works with every MCP client
One URL gives your tool Remote and Memory. You can switch each capability on or off from the dashboard.
Open Settings, then MCP, then Add Server, and paste the URL.
https://relay.flexx.dev/v1/mcpQuestions people ask first
Is my machine exposed to the internet?
No. Your machine only makes outbound WebSocket connections to the relay. Zero inbound ports are opened. The relay mediates all communication between AI agents and your machine.
What do I need to get started?
A free account and the flexx CLI on the machine you want to reach: install it with one command, run flexx setup and flexx connect, and that machine is reachable from anywhere. Memory needs only an account. To power Flexx Agent with your own models, add a provider API key in the dashboard.
What is MCP and how does flexx use it?
MCP (Model Context Protocol) is an open standard that lets AI agents interact with external tools. flexx turns your machine into an MCP server — AI agents like Cursor, Claude, and ChatGPT can read files, run commands, and edit code on your machine through the relay.
Which AI tools are supported?
Any tool that supports the Model Context Protocol (MCP) works with flexx. This includes Cursor, Claude Desktop, ChatGPT with MCP plugins, OpenAI Codex, and any custom MCP client.
Which platforms are supported?
flexx runs on macOS (Intel and Apple Silicon), Linux (x86_64 and ARM64), and Windows (x86_64 and ARM64). It's a single Go binary with zero dependencies.
How is authentication handled?
flexx uses OAuth 2.1 with PKCE for MCP connections. When an AI tool connects, it goes through a consent flow where you approve access and choose which machine to connect. You can revoke access anytime from the dashboard.
What tools can AI agents use on my machine?
AI agents can read and write files, execute shell commands, search codebases (via ripgrep), take screenshots (if Chrome is installed), list processes, and more. Available tools are auto-discovered based on what's installed.
Does it work behind a corporate firewall or NAT?
Yes. Because flexx only makes outbound connections, it works behind firewalls, NAT, and most corporate networks without any configuration. No VPN or port forwarding needed.
Is flexx free?
Yes. The free tier includes unlimited remotes, unlimited MCP tools, SSH access, and localhost tunneling. Managed and team tiers with additional features are coming soon.
Does flexx include agent memory?
Yes. flexx ships a graph-backed memory layer. Agents can store and recall decisions, preferences, and project context across sessions and tools — all through the same MCP endpoint.
Can I use flexx without installing anything?
Yes — for Memory. Create a free account and connect any MCP client; the memory tools need no install and no machine. Remote tools (files, shell, screenshots, tunnels) only become available after you install the flexx CLI on a machine and connect it. Flexx Agent can chat without a machine, but it needs a connected machine to act on one.
What is the Router, and do I have to use flexx's models?
No. Router is a bring-your-own-model gateway. Add an Anthropic, OpenAI, OpenRouter, Groq, or any OpenAI-compatible key in the dashboard, set per-key budgets, and use those models inside Flexx Agent — or point your own code at the router's OpenAI- and Anthropic-compatible endpoints.
What's the difference between using Remote/Memory directly and using Flexx Agent?
Remote, Router, and Memory are independent MCP servers you can plug into Cursor, Claude, or any client today. Flexx Agent is a unified assistant that already has all three connected, so you don't have to configure each one separately.
Install once. Reach that machine from anywhere.
$ curl -fsSL https://cli.flexx.dev/install.sh | shWindows PowerShell: irm https://cli.flexx.dev/install.ps1 | iex