Open Source Infrastructure Platform

Your Infrastructure.
Your Rules.

Deploy applications, LLMs, and MCP tools on your own hardware — as simple as making churros. No vendor lock-in, no surprise bills, full control over your data and infrastructure.

View on GitHub
https://portal.churrostack.com

Built for the tools you already use

Kubernetes
Docker
NVIDIA
Streamlit
vLLM
FastMCP
Istio
HuggingFace
Kubernetes
Docker
NVIDIA
Streamlit
vLLM
FastMCP
Istio
HuggingFace
Key Features

Everything you need to run your infrastructure

A complete platform for deploying, managing, and securing applications and AI workloads on your own hardware — without the complexity.

Private Hosting
Deploy on your own servers. Your data never leaves your network. The platform only handles secure communication through reverse SSH tunnels.
One-Click Deployments
Deploy Docker containers, Streamlit apps, or source code directly from Git. No DevOps expertise required.
LLM Hosting & Routing
Host models from HuggingFace with vLLM, or route requests through a unified OpenAI-compatible endpoint with fallback and usage tracking.
Workspace Mode
Give each user their own isolated deployment of any app — VSCode, development tools, or agentic workflows — all from a single configuration.
MCP & LLM Gallery
Centralize AI resources for your organization. Share models, MCP tools, and API keys securely with teams and external clients.
Enterprise Security
RBAC, OAuth/SSO, mTLS, network policies, and resource isolation built in. Every deployment is secured by default with Istio.
Environment Management
Slice your infrastructure into environments for teams, departments, or stages (DEV/QA/PROD) with dedicated resource quotas.
Remote Desktop
Run GUI applications like Playwright crawlers or browser-based tools with full remote desktop support via x11-vnc.
How It Works

From bare metal to production in minutes

Three simple steps to turn any server into a fully managed deployment platform.

Step 1

Connect Your Hardware

Install the ChurroStack agent on any server (even behind NAT or home networks). A reverse SSH tunnel securely connects it to the platform — no public IP or VPN required.

Step 2

Configure Environments

Slice your server into environments using Kubernetes namespaces. Set CPU, memory, and GPU quotas. Assign teams and users with fine-grained RBAC.

Step 3

Deploy & Manage

Launch applications, LLMs, or MCP tools through a simple browser wizard. The platform handles deployment, monitoring, logs, authentication, and scaling.

Cloud

portal.churrostack.com

Control Plane

Dashboard
Auth
Metrics
Proxy
RBAC
Orchestrator
SSH Tunnel

On-Premises

Your Private Network

Behind NAT / Firewall

Kubernetes Cluster

Env: Production

Streamlit
vLLM
MCP

Env: Development

VSCode
Docker
Resources64 CPU · 256 GB · 4× GPU

ChurroStack connects your private infrastructure to a managed control plane through secure reverse tunnels

Use Cases

Built for real-world scenarios

From AI infrastructure to internal tools, ChurroStack adapts to the way your organization actually works.

GPU-Powered AI
Got NVIDIA H100s collecting dust? Deploy vLLM models, encoders, rerankers, and AI document analyzers like Docling — without the Kubernetes headache.
Distributed Team Hardware
Turn spare hardware from team members' homes or offices into a unified deployment platform. Even machines behind ISP NAT work perfectly.
Secure Business Apps
Deploy internal Streamlit dashboards and business applications without exposing them to public cloud. Built-in authentication and monitoring included.
Automated Web Workflows
Build crawlers and RPA bots that connect to government portals or sites requiring human CAPTCHA verification. Full remote desktop support for browser-based interactions.
Exclusive Feature

Workspace Mode: One Config, Personal Spaces

A novel approach to multi-user deployments that gives every user their own isolated environment — without multiplying your configuration work.

Define your application configuration once, and ChurroStack automatically creates an isolated Kubernetes deployment for each user who accesses it. The same URL intelligently routes to the right deployment based on the logged-in user. Perfect for VSCode instances, agentic development tools, or any app that benefits from personal workspaces. Each user can securely manage their own API keys — like a personal Claude subscription — without exposing credentials to administrators or other users.

  • Isolated Kubernetes deployment per user
  • Single URL with smart routing
  • Secure API key management per user
  • Perfect for dev tools and AI workspaces

Single

Configuration

A

User A

Isolated Deployment

B

User B

Isolated Deployment

C

User C

Isolated Deployment

HuggingFace

OpenAI

vLLM

OpenRouter

Unified

ChurroStack Router

OpenAI-compatible API

Teams

Clients

Tools

AI Infrastructure

Your Enterprise AI Hub

Centralize your organization's AI infrastructure with a single platform for hosting, routing, and managing LLM access.

Host open-source models from HuggingFace on your own GPUs using vLLM, and expose them through an OpenAI-compatible API endpoint. The built-in router acts as a unified proxy — track usage per user and team, protect API keys from exposure, configure automatic fallback between LLM providers, and share access securely with internal teams or external clients using tools like ChatGPT, Claude, or Microsoft Copilot.

  • Host any HuggingFace model with vLLM
  • OpenAI-compatible unified endpoint
  • Usage tracking and billing per user/team
  • Automatic fallback between LLM providers
  • Share with ChatGPT, Claude, Microsoft Copilot
Pricing

Simple, transparent pricing

Plans that scale with your needs. Pay for the platform, not your compute.

Personal
For individual developers and small projects
$15/month
  • 1 Environment
  • 3 Services/Apps
  • 5 GB Traffic
  • 7-day Metrics & Logs
  • Community Support
  • No SLA
  • 1 User
  • Shared Infrastructure
Most Popular
Professional
For growing teams and production workloads
$75/month
  • 3 Environments
  • 15 Services/Apps
  • 25 GB Traffic
  • 7-day Metrics & Logs
  • Email Support
  • SLA Level C
  • 10 Users
  • Shared Infrastructure
Team
For departments and multi-team organizations
$500/month
  • 10 Environments
  • 100 Services/Apps
  • 200 GB Traffic
  • 7-day Metrics & Logs
  • Priority Email Support
  • SLA Level B
  • 100 Users
  • 1 Directory Sync
  • Shared Infrastructure
Enterprise
For large organizations — scales with your needs
Starting at$4,000/month
  • 50+ Environments
  • 1000+ Services/Apps
  • 1 TB+ Traffic
  • 90-day Metrics & Logs
  • PM + Phone Support
  • SLA Level A
  • Unlimited Users
  • 1+ Directory Sync
  • Dedicated/Isolated Infrastructure
  • Custom Domain

Additional bandwidth: $0.10/GB

FAQ

Frequently Asked Questions

Get Started

Ready to take control of your infrastructure?

Request a personalized demo and see how ChurroStack can transform the way your team deploys and manages applications.