01 Cloud models, local models, your hardware
Connect your models.
Share your GPUs.
Use one API.
Bring cloud models, local models, and your own GPU servers into one fleet. Use them yourself or make them available to others. As machines come online or go offline, Mycel routes new requests to available providers that meet your requirements.
Move or click to interact
02 Connect your hardware
Manage your own
private cloud.
Run your own Mycel Exchange to manage a private fleet of model servers for your team. Choose which machines join and give your apps one API to use them. Run a model server on your machine or GPU server, register with the Exchange operator, and connect the Mycel supplier agent. Your models can serve requests while connected and stop receiving new requests when you disconnect. Reconnect to make them available again—apps keep using the same API.
03 Available models
See what’s available
right now.
Browse models and prices across the fleet. Mycel matches each request to an available provider based on your requirements and price.
Loading models...
04 How it works
Send a request.
Get a response.
Your app sends an OpenAI-compatible request. Mycel selects an available provider, returns the response, and records cost and usage.
-
01 · Request
Your application
POST /v1/chat/completionsSends a standard request using its Mycel API key and optional end-user identifier.
-
02 · Route
Mycel Exchange
capability + price + availabilityMatches the request against live offers and selects an eligible supplier.
-
03 · Response
Selected model
stream + usage recordStreams tokens back through the same connection while Mycel records cost and usage.
05 For developers
Use your existing API.
Point your OpenAI-compatible requests at Mycel with your Mycel API key. Mycel handles routing and tracks cost and usage by app. Add an optional end-user ID to track usage by user.
- OpenAI-compatible chat completions
- Streaming and tool calling on supported models
- Cost and usage tracking by app and user
const response = await fetch(
"https://mycel.thefocus.ai/v1/chat/completions",
{
method: "POST",
headers: {
Authorization: `Bearer ${MYCEL_API_KEY}`,
"X-Mycel-End-User": user.id,
"Content-Type": "application/json"
},
body: JSON.stringify({
model: "deepseek/deepseek-v4-pro",
messages
})
}
); OpenAI-compatible endpoint
06 Your account
Loading your account…
Set up your first app.
Name your Client and first Application. We will issue an API key once; Mycel stores only its cryptographic hash.
Client account
—
01
Applications
02
Recent usage
07 Get started
Run models. Connect hardware.
Start making requests.
Create an account to get an API key, manage your apps, and view request history.