# RTK

43 Hacker News commenters recommended RTK and 44 criticized it, November 2025 to November 2026. It was recommended in 8 of 13 months; owners who said how long report a median of 0.04 years (4 owners).

**Verdict:** Wraps shell commands to trim output for AI coding agents, saving tokens for some users, though critics say savings are small and unmeasured.

## Why people recommend it

- Cuts command output tokens (6 comments): "on my Mac I've shaved off roughly 51k input tokens, and 23k output tokens, and saved an average of 3 seconds per command" ([source](https://news.ycombinator.com/item?id=48591848)) "tools like RTK, Headroom, caveman mode and others do reduce input and output tokens that need to be processed" ([source](https://news.ycombinator.com/item?id=48592020)) "It uses hooks to replace certain commands with optimized ones and helps reduce the number of input tokens" ([source](https://news.ycombinator.com/item?id=47593200))
- Reports visible token savings (4 comments): "I just looked at the RTK gain, and it seemed to do the job." ([source](https://news.ycombinator.com/item?id=50273879)) "That's significant token budget reclaimed." ([source](https://news.ycombinator.com/item?id=46974784)) "In 1 week. Numbers speak for themselves." ([source](https://news.ycombinator.com/item?id=46976200))
- Works with coding agents (4 comments): "It uses hooks to replace certain commands with optimized ones and helps reduce the number of input tokens" ([source](https://news.ycombinator.com/item?id=47593200)) "for agentic coding e.g. Claude Code I found this useful" ([source](https://news.ycombinator.com/item?id=48457806)) "I believe RTK would work well in that use case." ([source](https://news.ycombinator.com/item?id=49658301))

## What people warn about

- Savings small or unproven (6 comments): "I tested both for weeks and they didn't really saved that much money for Opus model." ([source](https://news.ycombinator.com/item?id=48082313)) "RTK saves between 2% and 7% on costs, it highly depends of your usage" ([source](https://news.ycombinator.com/item?id=50286877)) "I doubt it's saving much, but the quality of the work feels similar." ([source](https://news.ycombinator.com/item?id=48589242))
- Errors on unsupported commands (4 comments): "I just ran the steps to reproduce and got `Error: prettier produced no output` on rtk (0.42.2)" ([source](https://news.ycombinator.com/item?id=48589423)) "it gives an error message rather than sending the full output of the command instead" ([source](https://news.ycombinator.com/item?id=48173503)) "always uses RTK unless RTK falls apart running a tool it does not support" ([source](https://news.ycombinator.com/item?id=48173096))
- Confuses agents, causing loops or retries (5 comments): "some people see that changing how commands output stuff can confuse some models" ([source](https://news.ycombinator.com/item?id=47796178)) "my agent got stuck in a loop with a faulty RTK command and could not escape it since RTK hard overwrites anything automatically" ([source](https://news.ycombinator.com/item?id=48175808)) "I've watched agents go around in circles or use ridiculous workarounds after being confused by rtk output." ([source](https://news.ycombinator.com/item?id=48592975))
- Can add turns and cost (3 comments): "Sure your saved 70% tokens on that tool call, but nothing in the metrics says whether you ran 3 tool calls instead of 1." ([source](https://news.ycombinator.com/item?id=48591422)) "the number of turns to complete went up so the overall cost of the conversation was higher" ([source](https://news.ycombinator.com/item?id=48174550)) "agents on average perform *more* turns and use *more* tokens to accomplish one task" ([source](https://news.ycombinator.com/item?id=49592165))

## Alternatives hackers recommend

- [Sibyllinesoft Scribe](https://hackerapproved.com/p/scribe.md): 7 recommend
- [AgentReady](https://hackerapproved.com/p/agentready.md): 5 recommend

## Cite as

Hacker Approved, "RTK", https://hackerapproved.com/p/rtk/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).