Run local AI from any
HTTPS website.

Browsers block HTTP requests from HTTPS pages. Your local Ollama can't be reached. This 175KB proxy fixes it — with hardware-locked licensing and zero dependencies.

Get Started — $29/year How It Works
175 KB binary Zero dependencies Windows + Linux + macOS CORS enabled Hardware-locked

The Problem

HTTPS blocks your local AI

Modern web apps are served over HTTPS. Browsers enforce mixed content blocking — they refuse to make HTTP requests from HTTPS pages. Your Ollama server at http://localhost:11434 is invisible to any website served over HTTPS.

This means: gbrain, Open WebUI, any LLM playground hosted on a real domain cannot reach your local Ollama. Your AI models are stranded.

Without Ollama Proxy:

Browser
https://app.com
BLOCKED
Mixed content
Ollama
http://localhost:11434

With Ollama Proxy:

Browser
https://app.com
Ollama Proxy
localhost:11435
Ollama
localhost:11434

The Solution

Ollama Proxy sits between your browser and Ollama. It listens on a proxy port with proper CORS headers, forwards requests to Ollama, and returns responses with the headers browsers need. Your AI models become accessible from any HTTPS website.

Your data never leaves your machine. The proxy runs locally. No cloud, no external servers, no data collection.

Features

175 KB Binary

Pure native C. No Python, no Node.js, no Java, no runtime. Download, run, done. Smaller than a favicon.

Hardware-Locked License

License key is tied to your machine's CPU, disk, and hostname via SHA-256 fingerprint. Can't be shared or pirated.

CORS Headers

Automatically injects Access-Control-Allow-Origin: * headers so browsers allow cross-origin requests.

Streaming Support

Full support for Ollama's streaming chat completions. Token-by-token responses forwarded in real time.

Cross-Platform

Windows x64, Linux x64/ARM64, macOS x64/ARM64. One codebase, compiled natively for each platform.

Zero Dependencies

No OpenSSL, no Python, no npm. Just the OS network stack. Runs on any machine that can run Ollama.

How It Works

1

Get Your Machine ID

Download the proxy and run ollama-proxy --fingerprint. This generates a unique hardware identifier from your CPU, disk, and hostname. Copy this — you'll need it for checkout.

2

Purchase a License

Paste your machine fingerprint during checkout. We generate a license key signed with HMAC-SHA256, locked to your hardware. You receive it instantly by email.

3

Start the Proxy

Run ollama-proxy --license YOUR-KEY. The proxy validates your key offline (no phone-home), starts on port 11435, and begins forwarding to Ollama on 11434.

4

Use From Any HTTPS Site

Point your web app at http://localhost:11435 instead of 11434. The proxy adds CORS headers and forwards everything. Your local AI just works.

Pricing

One binary. One license. One machine. No subscriptions, no cloud fees, no data collection.

Annual

$29/yr
Renews annually
  • Full proxy functionality
  • All platform binaries (Win/Linux/Mac)
  • Hardware-locked to 1 machine
  • Free updates for 12 months
  • Email support
  • Transfer to new machine (1x/year)
Choose Annual

Team

$199/yr
Up to 10 machines
  • Everything in Lifetime
  • 10 machine licenses
  • Centralized license management
  • Slack/Discord support channel
  • Custom integration help
  • Invoice billing available
Choose Team

Download

Download the proxy binary for your platform. 175KB, zero dependencies, runs instantly.

💻

Windows x64

Windows 10/11, 64-bit

Download .exe

176 KB — ollama-proxy.exe

🐧

Linux x64

Ubuntu, Debian, RHEL, etc.

Download Binary

ollama-proxy-linux-x64

🍏

macOS

Intel or Apple Silicon

Download Binary

ollama-proxy-macos

Quick Start After Download

# 1. Get your machine fingerprint
$ ollama-proxy --fingerprint
# 2. Purchase a license (paste fingerprint at checkout below)
# 3. Start the proxy (local mode)
$ ollama-proxy --license YOUR-KEY
# 4. Or start in remote mode (for hosted gbrain)
$ ollama-proxy --license YOUR-KEY --bind 0.0.0.0 --api-key MY-SECRET

A license key is required. Purchase below, then enter it when starting the proxy.

Purchase

Your license key will be emailed within 60 seconds of payment. 30-day money-back guarantee. No questions asked.

FAQ

What is Ollama? +
Ollama is an open-source tool for running large language models (LLMs) locally on your machine. It runs models like Llama, Mistral, DeepSeek, and others. Ollama Proxy is a companion tool that makes Ollama accessible from HTTPS websites.
Why can't I just use Ollama directly? +
You can — if your web app is served over HTTP. But modern web apps use HTTPS, and browsers block HTTP requests from HTTPS pages (mixed content blocking). The proxy solves this by adding the required CORS headers and handling the protocol bridging.
Does the proxy send my data anywhere? +
No. The proxy runs entirely on your machine. It forwards requests from your browser to your local Ollama instance — both on localhost. No data leaves your network. The license validation is also offline — no phone-home.
Can I use one license on multiple machines? +
No — each license is locked to one machine's hardware fingerprint. If you need multiple machines, choose the Team plan ($199/yr for up to 10 machines). You can transfer a license to a new machine if you replace hardware.
What happens when my license expires? +
The proxy stops working. You'll see an "expired" message when you try to start it. Renew at any time to get a new key. Lifetime licenses never expire.
Is there a free trial? +
Not currently, but we offer a 30-day money-back guarantee. If the proxy doesn't work for your use case, email us and we'll refund you — no questions asked.
What web apps work with this? +
Any web app that calls the Ollama API. This includes gbrain, Open WebUI, text-generation-webui, LangChain web interfaces, custom ChatGPT-style frontends, and any app that uses the OpenAI-compatible API on localhost:11434.