real-time streaming
output appears as the task runs. no polling, no refresh. SSE-based streaming with automatic reconnection.
task
display long-running agent task output with real-time streaming, status indicators, tabbed views, and auto-scrolling logs.
pnpm dlx shadcn@latest add @inferencesh/taskTaskOutputWrapper handles everything — connects to the task, streams output in real-time, renders status indicators and tabbed views. compact mode for inline display inside tool invocations.
import { TaskOutputWrapper } from '@inference/ui/task' // renders a task by ID with live streaming <TaskOutputWrapper taskId="5rczn5xbw36dk9rr4h37n9sajz" /> // compact mode for inline display <TaskOutputWrapper taskId={taskId} compact />
color-coded status pills for the full lifecycle. queued, running, completed, failed — each renders with the right color and label automatically.
task is in the queue, waiting to start
task is executing, output streams in real-time
task finished, output and logs available
task errored, shows error message and logs
output, logs, and raw JSON in separate tabs. structured output fields render with appropriate formatting — images, captions, model info. auto-scrolling logs with pause on scroll-up.
features
output appears as the task runs. no polling, no refresh. SSE-based streaming with automatic reconnection.
output, logs, and raw JSON in separate tabs. switch between human-readable output and structured data.
logs scroll to bottom as new lines arrive. pause auto-scroll by scrolling up, resume by scrolling to bottom.
color-coded status pills with animated indicators. queued, running, completed, failed — at a glance.
structured output fields render with appropriate formatting — images, links, JSON, code blocks, markdown.
inline display for embedding task output inside tool invocations or other components.
start with the hosted platform. deploy your own when you're ready.
we use cookies
we use cookies to ensure you get the best experience on our website. for more information on how we use cookies, please see our cookie policy.
by clicking "accept", you agree to our use of cookies.
learn more.