Real-time system metrics for TermUI. CPU, memory, disk, network, processes, and HTTP health; available as raw function calls or as reactive hooks for JSX components.
npm install @termuijs/dataCall these functions directly for one-shot reads:
import { cpu, memory, disk, network, processes, system } from '@termuijs/data'
// CPU
console.log(cpu.percent) // 45.2
console.log(cpu.cores) // 8
// Memory
console.log(memory.used) // bytes
console.log(memory.total) // bytes
console.log(memory.percent) // 67.1
// Disk
console.log(disk.percent) // 52.3
// Top processes by CPU
const top = processes.top(5)
// [{ name: 'node', pid: 1234, cpu: 12.3, mem: 5.6 }, ...]
// Network throughput
console.log(network.rx) // bytes/sec received
console.log(network.tx) // bytes/sec transmitted
// System info (static)
console.log(system.hostname)
console.log(system.platform)
console.log(system.uptime)Use these inside JSX components. Each hook polls on an interval and triggers a re-render when data changes. All hooks clean up their intervals on component unmount.
import {
useCpu,
useMemory,
useDisk,
useNetwork,
useTopProcesses,
useSystemInfo,
useHttpHealth,
} from '@termuijs/data'
function CpuMonitor() {
const cpu = useCpu(500) // refresh every 500ms
return <gauge label="CPU" value={cpu.usage / 100} />
}
function SystemDashboard() {
const cpu = useCpu(1000)
const mem = useMemory(1000)
const disk = useDisk(5000)
const top = useTopProcesses(5, 2000)
const info = useSystemInfo() // fetched once; static
return (
<Box flexDirection="column">
<Text>Host: {info.hostname}</Text>
<Text>CPU: {cpu.usage.toFixed(1)}%</Text>
<Text>MEM: {(mem.used / mem.total * 100).toFixed(1)}%</Text>
</Box>
)
}import { useHttpHealth } from '@termuijs/data'
function HealthPanel() {
const checks = useHttpHealth(
['https://api.example.com/health', 'https://db.example.com/ping'],
5000, // check every 5 seconds
)
return (
<Box flexDirection="column">
{checks.map((c) => (
<Text key={c.url} color={c.status === 200 ? 'green' : 'red'}>
{c.url}: {c.status} ({c.latencyMs}ms)
</Text>
))}
</Box>
)
}| Hook | Default interval | Returns |
|---|---|---|
useCpu(ms?) |
1000ms | { usage, cores, model } |
useMemory(ms?) |
1000ms | { used, total, free, percent } |
useDisk(ms?) |
5000ms | { used, total, free, percent } |
useNetwork(ms?) |
1000ms | { rx, tx, interface } |
useTopProcesses(n, ms?) |
2000ms | Array<{ name, pid, cpu, mem }> |
useSystemInfo() |
once | { hostname, platform, arch, uptime } |
useHttpHealth(urls, ms?) |
5000ms | Array<{ url, status, latencyMs }> |
Raw collectors (cpu.percent, memory.used, etc.) do not clean up. Use them in one-shot scripts or outside component trees. Use hooks inside JSX components; they register cleanup on unmount automatically.
Full docs at www.termui.io/docs/data/overview.
MIT
@termuijs/data includes reactive hooks for fetching remote data inside
TermUI components. Each hook integrates with the TermUI render loop — updates
trigger re-renders automatically. No manual subscription management is needed.
The data-fetching hooks follow the same reactive model as the system-metric
hooks (useCpu, useMemory, etc.) already in this package. Each hook returns
a plain object whose fields update when new data arrives, and the component
re-renders only when the values it reads change.
Four primitives cover the most common patterns:
| Hook | Use case |
|---|---|
useFetch |
One-shot HTTP request on mount, with optional caching |
usePolling |
Repeatedly call an async function on a fixed interval |
useSSE |
Streaming updates over a Server-Sent Events connection |
invalidate |
Drop a cached response to force the next useFetch to re-fetch |
useFetch fires an HTTP GET request when the component mounts and exposes the
result reactively. Pass staleTime to cache the response in memory and skip
the network on subsequent mounts while the entry is still fresh.
import { useFetch } from '@termuijs/data'
import type { UseFetchOptions, UseFetchResult } from '@termuijs/data'
function PriceWidget() {
const options: UseFetchOptions = { staleTime: 30_000 } // cache 30 s
const result: UseFetchResult<{ price: number }> =
useFetch<{ price: number }>('https://api.example.com/price', options)
if (result.loading) return <Text>Loading…</Text>
if (result.error) return <Text color="red">{result.error.message}</Text>
return <Text>Price: {result.data?.price}</Text>
}UseFetchOptions fields:
| Field | Type | Description |
|---|---|---|
staleTime |
number |
How long (ms) a cached response stays fresh (default 0) |
UseFetchResult<T> fields:
| Field | Type | Description |
|---|---|---|
data |
T | null |
Parsed response body, or null while loading |
error |
Error | null |
Set if the request failed, otherwise null |
loading |
boolean |
true while the request is in flight |
usePolling repeatedly executes an async function on a fixed interval. Pass
any async function that returns data — it is not limited to HTTP requests.
import { usePolling } from '@termuijs/data'
import type { UsePollingResult } from '@termuijs/data'
function StockTicker() {
const result: UsePollingResult<{ symbol: string; bid: number }[]> =
usePolling(
() => fetch('https://api.example.com/quotes').then(r => r.json()),
5_000, // repeat every 5 seconds
)
if (result.loading) return <Text>Connecting…</Text>
return (
<Box flexDirection="column">
{result.data?.map((q) => (
<Text key={q.symbol}>{q.symbol}: {q.bid}</Text>
))}
{result.error && <Text color="red">{result.error.message}</Text>}
</Box>
)
}UsePollingResult<T> fields:
| Field | Type | Description |
|---|---|---|
data |
T | null |
Last successful return value, or null before first success |
error |
Error | null |
Last error, if any; polling continues regardless |
loading |
boolean |
true until the first execution completes |
useSSE opens a persistent Server-Sent Events connection and pushes each
received event into data. The connection closes automatically when the
component unmounts or url changes. An optional parse function lets you
deserialise the raw event string into a typed value.
import { useSSE } from '@termuijs/data'
import type { UseSSEResult } from '@termuijs/data'
interface LogLine {
ts: number
level: string
msg: string
}
function LiveLog() {
const result: UseSSEResult<LogLine> = useSSE<LogLine>(
'https://api.example.com/logs/stream',
(raw) => JSON.parse(raw) as LogLine,
)
if (result.loading) return <Text>Connecting…</Text>
return (
<Box flexDirection="column">
{result.error && <Text color="red">{result.error.message}</Text>}
{result.data && (
<Text>[{result.data.level}] {result.data.msg}</Text>
)}
</Box>
)
}UseSSEResult<T> fields:
| Field | Type | Description |
|---|---|---|
data |
T | null |
Most recent parsed event payload, or null before first event |
error |
Error | null |
Set if the connection failed, otherwise null |
loading |
boolean |
true until the first event is received |
useFetch caches responses in memory keyed by URL. A cached entry is reused
as long as it is younger than staleTime. Once the entry expires, the next
render re-fetches automatically.
To drop a cached entry early — for example after a write — call invalidate:
import { useFetch, invalidate } from '@termuijs/data'
function TodoList() {
const result = useFetch<{ id: number; title: string }[]>(
'/api/todos',
{ staleTime: 60_000 },
)
async function addTodo(title: string) {
await fetch('/api/todos', {
method: 'POST',
body: JSON.stringify({ title }),
})
// drop the cached list so the next render re-fetches
invalidate('/api/todos')
}
// …
}invalidate(key: string): void removes the cache entry for the given URL and
cancels any in-flight de-duplicated request for that key. Any mounted component
using the same URL will re-fetch on its next render. Calling invalidate with
a key that has no cached entry is a no-op.