headroom
Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Runssource: GitHubhomepagePythonApache-2.0commit 5d025f7a0387
Python, Apache-2.0 licensed. The project labels itself: agent, ai, anthropic, claude code, compression, context engineering, context window and cursor.
headroom runs. An Argusic agent installed it in 31 minutes and hit 4 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
Rust workspace (headroom-core, headroom-proxy, headroom-parity, headroom-simulators) builds from source and all cargo tests pass.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 100 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- No C compiler (gcc/cc) or binutils in container. Also no pip/venv for Python.15 minutes
- GCC fixed include headers (stdint.h) use #include_next which needs system headers in correct search order5 minutes
- C++ compiler needed by esaxx-rs crate (needs cc1plus)3 minutes
- libc.so linker script has hardcoded absolute paths (/usr/lib/x86_64-linux-gnu/libc_nonshared.a)3 minutes
- Install time
- 31 minutes
- Cold machine to finish
- 32 minutes
- Errors hit and fixed
- 4 hit, 4 fixed with no human help
- Model tokens used
- 501,722
- Exact commit tested
- 5d025f7a0387
- Container image
- sha256:4c3d41857be3
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- How the result was proved
- runs before 2026-09-07 did not record the command used to check the result
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
**Result:** `cargo check --workspace` builds and `cargo test --workspace` passes all 925+ tests acr… tokens used 501,722 ## Summary **Headroom** (context compression layer for AI agents) installed and tested successfully. **Blockers found and fixed:** 1. **Missing C/C++ toolchain** — Container had no `gcc`, `g++`, `binutils`, or system headers. Down… 2. **Hardcoded absolute paths in libc.so linker script** — The extracted `libc.so` referenced `/usr… 3. **Missing system headers for `#include_next`** — GCC's fixed `stdint.h` and C++ headers use `#in… 4. **Missing C++ compiler for `esaxx-rs` crate** — Downloaded and installed `g++-14-x86-64-linux-gn… 5. **Python unavailable** — No `pip`/`venv` in container, so the Python side and PyO3 `headroom-py`… **Result:** `cargo check --workspace` builds and `cargo test --workspace` passes all 925+ tests acr…
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Recovered from all 4 errors without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed Apache-2.0, as reported by its host.
What did not, or is not known
- Took 31 minutes to install, slower than the median of the 32 comparable projects Argusic has measured.
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
Of the 33 agent projects Argusic has installed and timed, headroom was the 32nd fastest to reach a running state, and 25 of 33 reached one at all.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
agentaianthropicclaude-codecompressioncontext-engineeringcontext-windowcursorfastapilangchainllmmcpopenaiprompt-engineeringproxypythonragtoken-optimizationtokenstypescript
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/headroom)Questions
- Does headroom run?
- Yes. headroom runs. Argusic installed and launched it on a clean machine in 31 minutes, hitting 4 errors on the way, and recorded the session.
- How did Argusic test headroom?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 5d025f7a0387. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test headroom?
- The run that produced this verdict cost $0.21: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does headroom take to install?
- 31 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does headroom need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing headroom?
- 4 things broke in the recorded run, and 4 were fixed without human help. Each one, and the time it cost, is listed on this page.
- How does headroom compare with the alternatives?
- Of the 33 agent projects Argusic has installed and timed, headroom was the 32nd fastest to reach a running state, and 25 of 33 reached one at all.
- Where is the evidence for headroom?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.