Apex — ML Platform

Jobs today▶

0

—

GPU hours used⏱

0.0

0 jobs active

Queue depth⊟

0

—

Success rate◉

—%

0 failed today

GPU utilisation — live

2s interval

60s

5m

GPU util %

CPU %

GPU util

—

CUDA kernels active

CPU util

—

— cores

Temp

—

Throttle @ 90°C

Submit job

GPU accelerated

Active jobs

View all

No jobs yet — submit one on the left.

Dev sessions

+ New

No active sessions.

Activity

Activity will appear here as jobs run.

All jobs

—

#	Name	Image	Status	GPU	Prio	Submitted	Duration	Actions
Loading…

Model registry

/workspace/models

Save model checkpoints to /workspace/models/<name>/
They will appear here once a scanner picks them up. (v0.2)

Team members

+ Invite

Loading team…

Plan

Team Plan

Self-hosted · unlimited GPU jobs

Active dev sessions

+ New session

No active sessions.

How dev sessions work

Each session launches a fully isolated environment with browser-native VS Code, pre-connected to your shared workspace. Click ↗ to open it instantly in a new tab — no setup, no SSH, no local install required.

Use the pre-built apex/code-server:python or apex/code-server:pytorch images, or bring your own.

GPU + CPU — live

2s SSE tick

GPU util %

CPU %

Current readings

GPU util

—

GPU temp

—

GPU power

—

VRAM used

—

VRAM total

—

CPU util

—

RAM used

—

RAM total

—

Docker images

—

Tag	Image ID	Size
Loading…

Tip

Apex reads images from your local image registry automatically — no manual push or sync required. Build your training image locally and it will appear here immediately. Pre-built apex/code-server:python and apex/code-server:pytorch images are available to get started quickly.

Audit log

Audit logging lands in v0.2. All API mutations will be recorded here.

Secrets

Encrypted secrets store lands in v0.2. For now, pass secrets via environment variables on the host machine.

Platform settings

Version

0.1.0

API base

—

Workspace path

~/apex-workspace

GPU

—

Config

~/.apex/config.json