Terminal answers in 300ms. No AI. No API. No waiting.
The command assistant that doesn't think. It just knows.
Why a terminal assistant matters right now
Developers are drowning in context-switching. The average engineer types a dozen commands a day, many of them half-remembered or needing a quick man-page lookup. Recent discussions on Hacker News (a 209-point thread with 45 comments) highlight a growing frustration: LLM-based terminal assistants like GitHub Copilot CLI or Warp AI are slow — often taking 1-3 seconds to respond — and they come with a recurring $10-20 monthly subscription. That's a tax on every keystroke, and a delay that breaks flow.
Existing solutions fall short. Copilot CLI is powerful but requires an API key, network round-trips, and a monthly fee. Warp AI locks you into its own terminal ecosystem. Traditional man pages are static and require you to translate documentation into your exact intent. And maintaining your own alias files is tedious and machine-specific. None of these offer a zero-cost, zero-latency, zero-learning-curve answer.
The timing is perfect. The HN thread itself proves the appetite for a "fast, no-LLM" alternative — many commenters explicitly complained about AI latency and cost. TERMy fills that vacuum: a purely local, rule-based engine that responds in under 300ms, with no API calls and no subscription. It's the tool that respects your flow and your wallet.
How It Works
Type your intent
Describe what you want in plain words — "find large files", "kill process on port 3000", "show git history for last 3 days". No special syntax, no flags to remember.
Get instant commands
TERMy's local rule engine matches your intent against a built-in library of 300+ hand-crafted rules. It returns 1-3 ready-to-run shell commands with a plain-language explanation for each.
Copy, run, or execute
Preview the command, copy it to your clipboard, or execute it directly with a confirmation prompt. Works in any shell — bash, zsh, fish — and respects your existing aliases.
What You Get
300ms response
Every query is matched locally against a pre-indexed rule set. No network round-trips, no token generation. You get answers in under 300ms — roughly 10x faster than typical LLM-based assistants.
$0 running cost
No API fees, no subscription, no hidden cloud compute. TERMy runs entirely on your machine. The open-source version is free forever, and even the Pro version is a one-time payment — no recurring charges.
300+ built-in rules
Covers the high-frequency commands you actually type: find, grep, kill, ps, network, git, docker, and more. Each rule is hand-written and tested, so you get accurate, copy-paste-ready commands without the fluff.