Dashboard Overview
The Fleet dashboard is mission control for your models. Everything you can do with the SDK and CLI you can also do here: upload models, kick off training runs, watch them live, deploy endpoints, and test inference. No terminal required.
Open it at fleethq.dev/dashboard.
Layout
The dashboard is organised into sections, reachable from the left sidebar:
| Section | What it's for |
|---|---|
| Overview | Live activity, running jobs, and recent models, jobs, and deployments at a glance. |
| Models | Your base models, fine-tuned outputs, and checkpoints. Train and deploy from here. |
| Datasets | Upload and manage training datasets. |
| Jobs | Full training-job history, with live logs and metrics. |
| Deployments | Live inference endpoints. Test them, copy their URLs, and shut them down. |
| API Keys | Create and revoke keys for the SDK and CLI. See Authentication. |
| Billing | Balance, transactions, storage costs, and auto-recharge. See Billing. |
| Settings | Account preferences and defaults. |
The Overview page
The landing page after sign-in. It surfaces what's happening right now:
- Activity on the GPUs: any job currently running, with its assigned GPU, status, and quick actions (metrics, logs, cancel).
- Recent activity: the latest jobs, models, and deployments across your account, newest first.
- Recent deployments: your active endpoints.
Your account balance is shown in the top bar. When it runs low, auto-recharge can top it up automatically.