zrok (Open-Source Tunnel) MCP. Expose local ports, manage tunnels via chat.
Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
zrok (Open-Source Tunnel) lets your AI agent expose local services to the world without complex firewalls or manual commands. It handles resource sharing for HTTP, TCP, and UDP ports.
You can manage environments, create public/private shares on demand, and monitor all tunnels directly through natural language prompts.
What your AI agents can do
Create share
Creates a new public or private tunnel share for your local service.
Delete share
Deletes an active share and shuts down the associated network tunnel.
Disable environment
Takes a registered environment offline, preventing new resources from being shared through it.
You tell the agent which local port to expose, and it creates a public or private tunnel for HTTP, TCP, or UDP traffic.
The agent can enable or disable entire environments, allowing you to register multiple machines or containers for resource sharing.
You retrieve your account status using get_account to see current usage counts against defined limits.
The agent runs list_shares to give you a list of every active tunnel and its associated details.
You tell the agent to delete a specific share using delete_share, immediately shutting down the exposed resource.
Ask AI about this MCP
Supported MCP Clients
Waiting for input…
zrok (Open-Source Tunnel): 10 Tools for Network Sharing
Manage resource sharing, tunnel creation, environment state, and account details using these ten specialized tools.
019e5d6ccreate share
Creates a new public or private tunnel share for your local service.
019e5d6cdelete share
Deletes an active share and shuts down the associated network tunnel.
019e5d6cdisable environment
Takes a registered environment offline, preventing new resources from being shared through it.
019e5d6cenable environment
Activates an environment, allowing machines or containers to register and share resources again.
019e5d6cget account
Retrieves your account status details, including usage limits for environments and shares.
019e5d6cget share
Gets detailed information about a specific existing share or tunnel.
019e5d6clist environments
Lists all the currently enabled development environments registered to your account.
019e5d6clist shares
Displays a list of every active, running share and tunnel on your account.
019e5d6clogin account
Authenticates the connection by exchanging credentials for an active zrok access token.
019e5d6cregister account
Sets up a brand new zrok account and obtains initial authorization tokens.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with zrok (Open-Source Tunnel), then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
zrok lets your AI agent expose local resources to the world. You don't have to deal with firewalls or manually running networking commands anymore; your agent handles all the messy stuff in a simple conversation.
To get started, you first need access. Your agent can run register_account if it’s brand new to zrok, setting up your account and grabbing your initial authorization tokens. If you're already set up, running login_account authenticates the connection by exchanging credentials for an active zrok access token. After that, you've got the keys to the kingdom.
Sharing Local Resources
You can expose any local port—HTTP, TCP, or UDP—by using create_share. You just tell your agent what service needs a tunnel and whether it should be public or private. The agent handles building that secure network path for you. For instance, if you run create_share for a web server on port 8080, the AI client instantly routes that service to a usable URL without you touching a single firewall rule.
Need to check what’s running? You can use list_shares, which gives you a full rundown of every active tunnel and its associated details. If you want specific info on one share—say, the status or destination of an existing tunnel—you run get_share and tell it exactly which resource you're checking.
When you're done with a service, don’t leave tunnels hanging. You shut them down immediately by calling delete_share, which removes the active share and kills the network connection associated with that tunnel. It’s clean, quick work.
Managing Environments
The server lets you manage multiple development environments so you can keep your resources organized. If you need to make sure no one accidentally uses an old setup, you run disable_environment on a specific environment name; this takes that entire environment offline and stops any new resources from being shared through it.
To bring a workspace back online, you use enable_environment. This reactivates the whole zone, letting machines or containers register and share their resources again. You can see all the available workspaces by listing them with list_environments.
Monitoring and Account Status
You keep tabs on everything your agent is doing. Running list_shares shows you every single active tunnel running right now, letting you know exactly what's exposed to the outside world. For a comprehensive status check—checking usage counts against defined limits or seeing how much bandwidth you’ve used—you run get_account. This gives you your current account health report.
If you need to manage which environments are even available for sharing, you can use list_environments to pull up the names of all currently registered development areas. Everything is contained in these simple commands: setting up accounts via register_account, authenticating with login_account, creating tunnels with create_share, monitoring everything with get_share and list_shares, controlling the scope with environment tools, and cleaning up with delete_share.
It’s that direct. No fluff, no manual CLI commands.
How zrok (Open-Source Tunnel) MCP Works
- 1 First, you authenticate your zrok Account Token via the agent's
login_accounttool. - 2 Next, you tell the agent to create a share (e.g., 'create a public share for port 80').
- 3 The system returns the live public URL and share token, giving you immediate network access.
The bottom line is that your AI client handles all the networking commands behind the scenes, letting you focus on development.
Who Is zrok (Open-Source Tunnel) MCP For?
This is for developers and platform engineers who are sick of manually running CLI commands just to share a local dev server. You're the Ops Engineer staring at a firewall dashboard late at night, wishing you could just type 'share port 3000' into your chat window.
You need to quickly expose a new API endpoint running on localhost:5000 for testing by a frontend team member without writing any firewall rules.
You manage multiple remote environments and need the agent to list all active tunnels (list_shares) across different development instances.
You use private sharing modes to grant temporary, audited access to an internal resource only for a specific team member.
What Changes When You Connect
- Manage resource sharing instantly. You don't need to open a terminal and run
create_shareevery time; just ask your agent to create the tunnel for you. - Stay within compliance boundaries. Before starting, use
get_accountto check your usage limits, so you never accidentally exceed your allocated environments or shares. - Full lifecycle control. Need to shut down a test server? Use
delete_share. It's instant access revocation with zero cleanup effort on your part. - Control development context. You can use
list_environmentsandenable_environmentto make sure the agent is pointing at the right machine or container group. - Zero-friction setup. Forget complex firewall rules. zrok routes traffic directly from your local host, making it dead simple for testing webhooks or local APIs.
Real-World Use Cases
The Webhook Test
A developer finishes building a payment webhook endpoint on localhost:80. Instead of asking an Ops team member to manually configure port forwarding, they just prompt their agent. The agent runs create_share for HTTP port 80 and returns a live public URL the client can hit immediately.
The Cleanup Job
After testing is done, the developer doesn't want lingering tunnels. They prompt the agent to list all shares using list_shares, identify the temporary tunnel ID, and run delete_share to ensure the resource is completely offline.
The Multi-Machine Setup
A DevOps engineer needs two separate dev environments running—one on Mac, one on Linux. They first use list_environments, then call enable_environment for both to ensure the agent can correctly manage sharing across multiple operating systems.
The Security Audit Check
A security team member needs to know what resources are currently exposed. They run get_account first to verify usage, then use list_shares to get a clear inventory of all active tunnels and their current status.
The Tradeoffs
Over-sharing access
A user runs create_share for an entire internal IP range instead of just the specific port, accidentally exposing sensitive network data.
→
Always specify the exact resource and scope. Use get_share first to verify the current share's details before changing it. If you need a temporary public tunnel, stick to create_share with granular port numbers.
Ignoring environment state
A user tries to create a new share but forgets which machine is active, resulting in the resource being routed through an old or disabled container.
→
Always check and manage your environments first. Run list_environments and use enable_environment before attempting any sharing action.
Leaving tunnels running
A temporary testing tunnel is left active after the job is done, wasting quota or posing a security risk.
→
Treat shares like disposable resources. As soon as you're finished, use delete_share to revoke access immediately and cleanly.
When It Fits, When It Doesn't
Use this server if your core need is securely exposing local machine ports (HTTP/TCP/UDP) over a network boundary without manual CLI work. You want the AI agent to handle the tunnel lifecycle—from create_share to delete_share. Don't use it if you are trying to transfer files or read data from a remote database; those require different types of connections.
If your main goal is managing multiple, distinct compute instances (e.g., dev machines vs. staging containers), then focus on the environment tools: list_environments, enable_environment, and disable_environment. If you just need to check usage quotas, get_account is your go-to tool.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by zrok. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Sharing a local service shouldn't take three dashboards and an Ops call.
Before zrok, sharing anything—even a simple dev server running on port 3000—meant logging into a portal, finding the correct IP range, submitting a ticket, waiting for human approval, and then manually configuring firewall rules. It was slow, it required multiple hands, and you always wondered if they remembered to turn it off.
Now, your AI agent handles all of that complexity. You simply ask it to 'share my local web server on port 3000.' The agent runs `create_share`, generates the public URL instantly, and gives you access in seconds. It's done.
zrok (Open-Source Tunnel) MCP Server: Manage Tunnels from Chat
Gone are the manual steps of checking tunnel status via a dedicated CLI, running `list_shares` to track names, and then manually hitting the portal to terminate access. The agent handles state management for you.
You interact with it like talking to a teammate: 'List all active shares,' or 'Delete that one on port 80.' It's immediate, conversational network administration.
Common Questions About zrok (Open-Source Tunnel) MCP
How do I start using the zrok (Open-Source Tunnel) MCP Server? +
You must first authenticate your connection. Use the login_account tool to exchange credentials for an active token before any other commands will work.
What difference is there between `list_shares` and `get_share` in zrok (Open-Source Tunnel) MCP Server? +
list_shares gives you a summary of every active tunnel on your account. Use get_share when you know the specific share name or ID and need all its detailed metadata.
Can I use zrok (Open-Source Tunnel) MCP Server to expose UDP traffic? +
Yes, it supports UDP. When calling create_share, you simply specify the protocol as UDP along with the required port number.
How do I stop a share using zrok (Open-Source Tunnel) MCP Server? +
Use the delete_share tool. This immediately terminates the tunnel and revokes public access to that specific service.
When should I use the `login_account` tool with zrok (Open-Source Tunnel) MCP Server? +
You must run this first to authenticate and get a fresh token. This refreshes your credentials, ensuring you don't have to manually re-authenticate when running scheduled tasks or developing applications.
How does `disable_environment` affect my active shares in zrok (Open-Source Tunnel) MCP Server? +
Disabling an environment immediately revokes all associated tunnels. Any share created within that specific environment becomes inaccessible, serving as a quick security cutoff switch when you're done testing.
What information does the `get_account` tool provide about my zrok (Open-Source Tunnel) MCP Server usage? +
It gives you a complete overview of your account limits and current consumption metrics. You can track exactly how many environments or shares you've allocated against your subscription quota.
Before running `delete_share`, how do I confirm the settings using `get_share` in zrok (Open-Source Tunnel) MCP Server? +
The get_share tool shows you all parameters, including whether it's public or private and its current protocol type. This confirmation prevents accidental deletion of critical tunnels.
How do I create a public tunnel for my local web server? +
Use the create_share tool. Specify the backend (e.g., http://localhost:8080), set share_mode to 'public', and backend_mode to 'proxy'. The agent will return the public URL.
Can I see all my currently active tunnels across all environments? +
Yes! Run the list_shares tool. It will retrieve a list of all active shares associated with your account, including their tokens and backend configurations.
How do I stop an environment from being active in zrok? +
You can use the disable_environment tool. This will remove the environment registration from the zrok controller, effectively disconnecting that machine or container.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Aptabase (Mobile Analytics)
Send telemetry and custom events to Aptabase directly from your AI agent to track app usage and performance.
T-Test Statistics Engine
Run exact Student's, Welch's, and Paired t-tests local. Get CPU-guaranteed p-values instead of LLM-hallucinated guesses.
ScraperAPI
Equip your AI agent with proxy rotation and headless browsers to extract HTML, Google SERPs, and Amazon data at scale.
You might also like
Fireflies.ai
Transcribe meetings, analyze conversation metrics, and query transcripts via AI agents with Fireflies.ai.
Colossyan
Automate AI video creation via Colossyan — generate videos from templates, manage avatars, and track rendering jobs directly from any AI agent.
Viral Loops
Manage referral campaigns, participants, milestones, and rewards via Viral Loops API.