Zsh is the shell; Oh My Zsh is an optional framework built on top of it. You can use Zsh without Oh My Zsh, and if your configuration loads more than you need, trimming or replacing that setup may help. Cason Adams describes zload as a plugin manager for loading selected plugins—including Oh My Zsh plugins—and deferring some work. His reported sub-millisecond figure is an author-reported warm-start result, not an independently verified measure of time to a usable prompt.
What is zload, and what problem is it meant to solve?
Oh My Zsh is not Zsh itself. It is a framework that organizes configuration and plugins for the Zsh shell; its FAQ notes that Zsh can be used without Oh My Zsh (Oh My Zsh FAQ). Replacing the framework is therefore an option, not a requirement for using Zsh.
As an Amazon Associate I earn from qualifying purchases.
In his article, Cason Adams presents zload as a plugin manager for people who want selected plugins without loading the broader Oh My Zsh framework. He describes support for Oh My Zsh plugin paths using an omz: prefix, ways to defer selected work until after the prompt appears or until a command is invoked, and compiling initialization output or plugin files into Zsh wordcode. Those are the author’s product descriptions; the available material does not independently establish their behavior or performance.
Why can an Oh My Zsh setup feel slow?
The framework alone does not determine startup time. A configured shell can spend time launching subprocesses during initialization, parsing shell scripts, and loading plugins or other framework features the user does not need. Adams identifies these as sources of overhead and positions selective loading as a way to avoid some of that work. That does not mean every Oh My Zsh installation is slow: the result depends on the particular configuration, plugins, machine, and what “startup” measures.
#1 Best Overall
Deferring work can also change when its cost is paid rather than remove it. If initialization is postponed until after the prompt or until a command runs, the prompt may appear sooner while a later action still takes time. A useful comparison should therefore consider both initial responsiveness and any delay introduced later.
What do the zload benchmark numbers actually say?
Adams’s comparison article reports warm-start overhead of 0.45 ms for zload with 10 active plugins and 185.00 ms for the Oh My Zsh framework, in a table labeled macOS (Apple Silicon). It also claims zero subprocess forks for zload under the described interactive startup. These are the author’s reported results, not independently reproduced general benchmarks. The reviewed article does not establish a complete repeatable test protocol or show that the compared configurations offer equivalent features.
| Reported result | Context stated in the article | How to read it |
|---|---|---|
| 0.45 ms | zload with 10 active plugins; warm startup; macOS (Apple Silicon), as labeled in the comparison table; reported by Cason Adams in 2026 (the result is indexed with a Sep 29, 2026 date) | Author-reported figure; the available sources do not independently reproduce it or establish that it is first-prompt latency. |
| 185.00 ms | Oh My Zsh framework; same comparison table and stated platform; reported by Cason Adams in 2026 (the result is indexed with a Sep 29, 2026 date) | Author-reported comparison figure; the article does not establish equivalent enabled features or a repeatable protocol. |
The article’s phrase “Warm Startup < 0.5 ms” should not be read as a promise that a terminal will show a usable prompt in under half a millisecond. Warm-start overhead and first-prompt latency are not necessarily the same measurement. To establish a fair comparison, a benchmark would need to state the exact metric, hardware and operating system, shell and plugin versions, cold or warm state, enabled features, and run count.
Why does the startup metric matter?
Interactive responsiveness has several parts: how long it takes for the first prompt to appear, how long until the first command can run, the delay when entering commands, and process exit time. The zsh-bench documentation distinguishes these measurements and cautions that time zsh -lic "exit" measures command execution time, which can differ from what an interactive user experiences.
Its examples illustrate why one number can mislead: an agnoster configuration is shown with 32 ms first-prompt lag and 2 ms exit time, while a powerlevel10k configuration is shown with 2 ms first-prompt lag and 6 ms exit time. These examples come from separate configurations and are not a controlled comparison with each other or with zload. They do show that exit time alone cannot tell you how soon a prompt becomes usable.
How can you measure and improve your own Zsh startup?
Start by profiling the configuration you already have. The Z-Shell community profiling guide documents use of zsh/zprof and guarded execution tracing. Make one change at a time, launch a fresh Zsh process for each comparison, and focus on the delay you actually want to reduce.
- Profile with zprof. Add
zmodload zsh/zprofnear the start of~/.zshrc, then addzprofafter the configuration lines you want to measure. Start a fresh Zsh process and inspect the report for expensive functions. - Trace initialization when the profile is not enough. The guide documents guarded execution tracing with
PROFILE_STARTUP=true zsh -i. Use the output to locate which startup commands are taking time; tracing can produce substantial output, so keep it targeted. - Test a fresh interactive shell after each change. The guide recommends a new Zsh process for each comparison. Keep the shell version, machine, plugins, and other conditions consistent, and distinguish first-prompt or first-command delay from process exit time.
- Remove or defer work deliberately. Disable plugins you do not use, or investigate whether expensive initialization can be delayed without making the commands or prompt you rely on feel slower. If considering zload, verify its current installation and usage details from its own project materials before changing your configuration; the figures above alone are not sufficient evidence of a performance gain on your machine.
Profiling can tell you whether a plugin, theme, or initialization command is the bottleneck in your setup. If the measured cost comes from work you need, the practical choice may be to accept it, replace that component, or defer it—not simply to change plugin managers.
Is zload faster than Oh My Zsh?
Adams reports much lower warm-start overhead for zload in his stated comparison, but the available evidence does not independently verify the result, establish equivalent configurations, or show that the figure represents time to the first usable prompt. It is not enough to conclude that zload is faster for every user. Profile your own interactive shell and compare the same user-facing metric before deciding whether a different loading approach solves your problem.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
