Dev tools

hqtui builds terminal dashboards in TypeScript and publishes what each frame costs

September 5, 2026 at 5:05 PM PT

The GitHub card for the profullstack/hqtui repository

Image: GitHub

Why it mattersA monitoring TUI you can render to plain text in a test is one you can assert on in CI, which is the difference between a dashboard you trust and one you eyeball.

hqtui is a TypeScript library for building terminal dashboards of the kind btop and htop produce. It has no runtime dependencies, ships under MIT, and the GitHub API showed 202 stars on 5 September 2026 against a repository created on 30 August, so that is six days. The most recent tagged release, v0.1.11, went out on 4 September.

The same code runs on Bun 1.1 or newer, Node 22.6 or newer, and Deno 2. The README says all three run the library and the demo unchanged, and that it is tested on Linux, macOS and Windows Terminal.

It publishes the frame cost, with the method attached

Most rendering libraries describe themselves as fast. This one gives figures and says where they came from: a 160 by 50 screen, which is 8,000 cells, on Bun 1.4 and Linux x64.

An unchanged frame takes 0.068ms and writes no output at all. A frame where 1 percent of the screen changed takes 0.140ms and writes 627 bytes. At 10 percent changed it is 0.291ms and 2,639 bytes. A full redraw of every cell is 1.409ms and 8,341 bytes. A six-panel dashboard, counting layout, widgets and the diff, renders in 0.425ms.

Those are the project's own measurements on its own machine, so they are a shape rather than a guarantee. The shape is the point: the renderer diffs and writes only what changed, and the byte counts show the saving directly. On a dashboard polling once a second over SSH, the difference between writing 627 bytes and 8,341 bytes per frame is the difference between a readable display and a flickering one.

Rendering to text is the part that changes a workflow

The testing utilities render a full interface without a terminal attached. Four functions produce plain text, a screen buffer, ANSI output, or HTML, with no TTY required.

That turns a class of code that normally goes untested into ordinary code. A monitoring view is usually verified by running it and looking at it, which means layout regressions ship. Rendering the same view to a string in a test and asserting on it makes the dashboard something CI can check, the same as any other output.

The widget set is wide enough to matter: panels, tables, trees, log viewers, meters, gauges, donuts, sparklines, multi-series graphs, histograms, heat bars, tabs, status bars, and the usual form controls including a command palette and modals. Colour handles 24-bit truecolor with automatic quantisation down to 256 and 16 colours, plus NO_COLOR, monochrome and high-contrast modes. On limited terminals the README says it degrades to no mouse, quantised colour and ASCII in place of Braille.

At v0.1.11 and six days old, the version number is the honest signal here: the API will move. What is worth taking now is the benchmark table itself. A library that tells you a full redraw costs 1.409ms and 8,341 bytes has given you enough to decide whether it fits your refresh rate, and that is more useful than any claim about being lightweight. Anyone shipping an internal tool with a live view of agent runs, queue depth or build status has a real reason to look.

Source

profullstack/hqtui, GitHub. Release v0.1.11 published 4 September 2026; star count from the GitHub API, checked 5 September 2026.

Source: GitHub

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

More from AI News

Lody shares a running coding agent session with the rest of your team

Lody connects any ACP-compatible coding agent to a shared workspace so a teammate can open the same live session from desktop, phone or web, and it reached 886 stars in under a month.

Dev tools

Codenotch puts your Claude, Cursor and Codex usage limits on the edge of the screen

Codenotch is a macOS app that pins how much of each coding agent's session limit you have used to the edge of the screen, and it reached 196 stars within hours of being published.

Dev tools

Ponytail makes a coding agent check seven things before it writes any new code

Ponytail is an MIT-licensed rule set that makes a coding agent work through a seven-step ladder of cheaper options before it writes new code, and it has reached 127,785 stars on GitHub.

Dev tools