Headroom
Context optimization layer in AI tools
Free open-source layer that compresses tool outputs before they reach the LLM, cutting token use. Critics say savings are unproven and it can break prompt caching.
9 people recommended Headroom in Hacker News comments, 4 had mixed feelings and 5 criticized it, November 2025 to October 2026.
Summary written by Claude from 30 comments. It was written before the November 2026 recount, and 24 of those comments still count.
18 people gave an opinion
- recommend it
- 9
- mixed feelings
- 4
- warn against or regret it
- 5
The record
Recommended in 4 of 12 months
In 7 different threads.
- November 2025: not recommended
- December 2025: not recommended
- January 2026: recommended
- February 2026: not recommended
- March 2026: not recommended
- April 2026: recommended
- May 2026: not recommended
- June 2026: recommended
- July 2026: not recommended
- August 2026: recommended
- September 2026: not recommended
- October 2026: not recommended
Why it doesn’t carry the badge
- 9 people recommend itNeeds 10 or more, not met.
- 1.8 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 4 of 12 monthsNeeds 4 or more, met.
- Recommended in 7 threadsNeeds 5 or more, met.
- The biggest thread supplies 33% of its recommendersNeeds half or less, met.
The same rules apply to every product, and the badge can’t be bought. Mixed feelings count toward neither side. How the badge works
What people say about it
Reasons grouped from a sample of 23 of the 28 opinions about it, picked in proportion to how many people recommend it, have mixed feelings and criticize it. Each mark is one comment in the sample. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Cuts token use and cost
“tools like RTK, Headroom, caveman mode and others do reduce input and output tokens that need to be processed”
Comment from 18 Jun 2026 in “The Token Compression Illusion: Why I'm Skeptical of RTK” “setting up a combination of both tools could reduce token consumption by 50%”
Comment from 19 Jun 2026 in “The Token Compression Illusion: Why I'm Skeptical of RTK” “the real tool you want to reduce token usage is headroom”
Comment from 16 Apr 2026 in “Claude Opus 4.7” Transparent benchmarks and algorithms
“They provide accuracy benchmarks in the repo and are transparent about the compression algorithms used for the different output types.”
Comment from 19 Jun 2026 in “The Token Compression Illusion: Why I'm Skeptical of RTK” “I like headroom’s approch a lot better. More transparent.”
Comment from 19 Jun 2026 in “The Token Compression Illusion: Why I'm Skeptical of RTK” “I applaud the benchmarking though.”
Comment from 19 Jun 2026 in “The Token Compression Illusion: Why I'm Skeptical of RTK” Keeps needed data, reversible
“compresses tool outputs (schema-preserving subset of JSON arrays; keeps errors/anomalies/top/relevant”
Comment from 13 Jan 2026 in “Headroom – context optimization layer for tool-using agents” “The goal here was: reduce tokens without breaking the contract.”
Comment from 13 Jan 2026 in “Headroom – context optimization layer for tool-using agents” Works with Claude Code and Cursor
“you've made it a joy to use! I've just made the v0.0.1 of an adapter to run headroom as a hook”
Comment from 12 Jun 2026 in “Headroom compresses everything your AI agent reads before it reaches the LLM”
What people warn about
Each mark is one comment
Token savings unproven or negligible
“with some limited "success" (for some definition of success”
Comment from 14 Sep 2026 in “Notes on gotchas while migrating 35kb preprompts from Opus to self-hosted Ollama” “Benchmarks done on Headroom and RTK show that neither result in real savings.”
Comment from 11 Sep 2026 in “RTK reports token savings, but our cost benchmarks disagree” “the rest of the efforts didn’t move the needle much”
Comment from 12 Apr 2026 in “Pro Max 5x Quota Exhausted in 1.5 Hours Despite Moderate Usage” - Comment from 27 Apr 2026 in “8v: One CLI for you and your AI agent. Up to 66% fewer tokens”
Can raise conversation cost
“the number of turns to complete went up so the overall cost of the conversation was higher”
Comment from 18 May 2026 in “Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep” “headroom violates cache frequently, nullifying any potential savings it could have”
Comment from 11 Sep 2026 in “RTK reports token savings, but our cost benchmarks disagree” Hard to configure
“That was a real PITA to configure for local-only”
Comment from 22 Sep 2026 in “Frontier AI on Your Own Hardware” Lacks transparency about how it works
“The lack of transparency here is really concerning.”
Comment from 17 Apr 2026 in “Claude Opus 4.7”
What people said
23 of the 28 opinions about it, picked in proportion to the three counts, newest first. Each links to the original comment.
“That was a real PITA to configure for local-only”
“with some limited "success" (for some definition of success”
“headroom violates cache frequently, nullifying any potential savings it could have”
“Benchmarks done on Headroom and RTK show that neither result in real savings.”
“it came back very clear that off was best”
“It runs elaborate subsystems towards a noble objective of saving tokens.”
“I applaud the benchmarking though.”
“setting up a combination of both tools could reduce token consumption by 50%”
“I like headroom’s approch a lot better. More transparent.”
“They provide accuracy benchmarks in the repo and are transparent about the compression algorithms used for the different output types.”
“tools like RTK, Headroom, caveman mode and others do reduce input and output tokens that need to be processed”
“you've made it a joy to use! I've just made the v0.0.1 of an adapter to run headroom as a hook”
“Headroom remains OSS and free to use”
“the number of turns to complete went up so the overall cost of the conversation was higher”
“I didn't notice any difference.”
“quickly decided to uninstall when I realized how invasive it is and requires a subscription”
“The lack of transparency here is really concerning.”
“You don’t notice it’s there.”
“Headroom looks great for client-side trimming”
“the real tool you want to reduce token usage is headroom”
“the rest of the efforts didn’t move the needle much”
“The goal here was: reduce tokens without breaking the contract.”
“compresses tool outputs (schema-preserving subset of JSON arrays; keeps errors/anomalies/top/relevant”
Counts come from 1,791,911 Hacker News comments labeled by Claude, November 2025 to October 2026, with data through October 31, 2026. Counts updated 2 Nov 2026. How we count. Report a mistake.