Help · Using Mealissa

Managing nodes & dashboard

Everything you run lives in the cloud console — nodes, groups, licenses, and a full usage dashboard.

Opening the console

Sign in with Google at mealissa.com/cloud. The console has three views, switched from the sidebar: Dashboard, Licenses, and Usage.

Dashboard — your groups and nodes

The dashboard shows every group you run and the nodes inside it, refreshing automatically (the online dot and countdown at the top show the refresh cycle). For each node you'll see:

  • Node ID — the identifier chosen (or generated) at the agent's first-run prompt.
  • Model badge — the Ollama model the node serves, with its parameter count (e.g. qwen2.5:4b · 4.0B).
  • Endpoints — the node's public and local ip:port.
  • Node JSON — a copy button for the node's full record, handy when scripting against the API.

A node appears here the moment its agent registers and disappears when the agent disconnects. Agents send a heartbeat every 30 seconds and reconnect automatically if the link drops — see Troubleshooting if a node you expect is missing.

Playground

When at least one node is online, the dashboard shows a playground: pick a model, type a prompt, and the answer comes back from your hardware — a zero-setup way to confirm end-to-end that everything works.

Hole-punch sessions

The dashboard also lists active hole-punch sessions — direct node-to-node connection attempts brokered by the cloud (used for peer-to-peer scenarios started with the agent's connect command). For most API users this table simply stays empty.

Licenses

The Licenses view lists your keys with tier, status, node limit, and expiry, plus a copy button for each key and the Request trial license button. Full walkthrough: Getting a license.

Usage

The Usage view is your traffic dashboard. Every API request through your license is logged with its model, node, latency, and status. You get:

  • KPI cards — totals for the selected range.
  • Range tabs — switch between the last 24 hours, 7 days, and 30 days.
  • Breakdowns — requests by model, by license key, and by node.
  • Recent requests — the latest calls with latency and status (ok / timeout).
  • Requests over time — a time-series chart of your traffic.
Note: token counts in usage data currently report 0 — volume, latency, and status are the metrics to watch.

Running more nodes

Repeat the install on each additional machine and enter the same group name at the first-run prompt. Group API calls (/g/<group>/...) are answered by any online node in the group, so adding machines adds capacity with no client-side changes. One node per group can be marked as the default (the Default prompt or the --default flag).

How many nodes one license may register is set by its tier — the limit is shown on your license card. See the licensing plans for tier details.

Next steps

See your fleet in one place

Nodes, licenses, and usage — live in the cloud console.

Open the console Add a node →