Signed in as:

Radix Core (SaaS Preview)

Intelligent GPU scheduling with adaptive control plane

Sign in to access the dashboard and create your API key.

Radix Core — Secure LLM Workflow Orchestration

Real-time monitoring and auditable execution for HPC research clusters

Cluster Overview

Real-time GPU utilization, queue depth, and cluster health metrics

GPU Utilization
—
Queue Depth
—
Running Jobs
—
Radix Score
—

🎯 Control Plane Advice

Loading...

GPU Utilization & Queue

Cluster Health Score

Live updates every 5 seconds • Cluster: vault-hub-demo

LLM Workflows

Active inference workflows with throughput, latency, and token metrics

🤖

No LLM workflows observed yet

Once your agent starts sending LLM telemetry, your workflows will appear here.

Workflow Latency Trends

Pipelines — Design, Test, Deploy

Design and test reusable AI workflows that call your registered models

Pipelines orchestrate one or more model calls into a reusable workflow.

🔗

No pipelines yet — create your first pipeline to get started

Pipelines let you chain model calls, retrieval, and aggregation into reusable workflows.

Pipeline:

Definition

🧪 Sandbox — Test Pipeline

Results

Steps
Outputs

New Pipeline

Step Configuration

Use {input} as a placeholder for the pipeline input

Run Pipeline:

Results

Models — Register & Manage

Bring your own models: HTTP endpoints, containers, or external clusters

🤖

No models registered yet

Register your first model to use in pipelines.

Register New Model

Clusters — External GPU Nodes

Connect external GPU clusters to run inference and training workloads

Name Cluster ID Type Status GPUs Last Heartbeat Actions
🖥️

No clusters registered yet

Register your first external cluster to run workloads on your own GPU infrastructure.

Register New Cluster

🎉 Cluster Registered Successfully!

⚠️ Important: Save your agent token now!

This token will only be shown once and cannot be retrieved later.


              
            

Experiments

Define experiments and launch hyperparameter searches on top of your existing pipelines and models.

Audit Log — LLM Workflow Runs

Complete, auditable history of every LLM inference run with model, tokens, latency, and status

Showing 0 runs
Time Workflow Model Version Tokens In Tokens Out Latency (ms) Score Status
📋

No audit records yet

LLM workflow runs will be logged here for full auditability.

🔐 Admin — User Management

Invite and manage users (admin-only)

Invite New User

Users

Email Roles Status Created
Loading...

📊 Usage Analytics

Loading...

User GPU Jobs CPU Jobs Last Run
Loading...

Note: Currently showing all-time usage. Time-window filtering (e.g., last 7 days) will be added in Phase 2.

⚙️ Cluster GPU Policies

Loading...

Update Policy

Note: GPU job limits prevent users from monopolizing cluster resources. CPU jobs are not limited in Phase 1.

⚙️ Settings & Subscription

Manage your Radix Core subscription and view tenant details

Tenant Information

Tenant ID: Loading...

Plan: Loading...

Status: Loading...

Core (Radix Core)

Core Plan: None

Licensed GPUs: 0

Trial: None

Studio (Radix Studio)

Studio Plan: None

User Limit: 0

Model Limit: 0

Pipeline Limit: 0

Team

Manage who has access to this Radix tenant. Seats are limited by your Studio plan.

Email Role Status
No team members yet.

New Experiment

Installation & Setup

Connect your cluster to Radix Core control plane

Step 1: Create API Key

Generate your API key to connect your cluster to the Radix control plane.

Your API Key

Step 2: Install on Your Cluster

Copy and paste this into a shell with kubectl + helm configured:

Step 3: Test the API

Try the scoring endpoint with curl:

HPC Metrics Dashboard

Real-time cluster performance indicators

GPU Utilization
—
Current load percentage
Queue Depth
—
Pending jobs
API Latency
—
Score calculation time

Score History

GPU Utilization Trend

Recommendations

Run a score calculation to see cluster-specific recommendations.

Live Score Playground

Test the scoring algorithm with different cluster metrics:

Terms of Service · Privacy Policy · Support © VaultScaler – Radix Core