colonizer
install

modules / token savers

headroom

squeezes large tool output, such as json, before the model reads it.

token saversoptional

status
optional ships with colonizer, off until you switch it on.
version
bundle 0.37.0-1
pinned
bundle 0.37.0-1, headroom-ai 0.37.0 · archive sha256 2aa548fb (x86_64) / 47ad96db (aarch64)
licence
apache-2.0
source
headroomlabs-ai/headroom ↗
docs
token savings

what it does for you

headroom is a proxy inside the colony that compacts large tool results, such as json, before the model reads them. in its smoke test a 400-row json log arrives 75% smaller, every error row kept.

what it needs

how to turn it on

settings → modules on the claude code settler: the headroom switch. off by default; switching it on starts the download.

how far it has been tested

it has run in a microvm set up like a colony and reached anthropic through the credential swap. no full claude code session has run through it, and it has not run in a colony yet.

how colonizer runs it

what it costs

221 mb for aarch64 and 243 mb for x86_64, downloaded once per release; 300 to 370 mb of the colony's memory (measured on arm64); 2.5 to 7 seconds to start.

← all token savers · every module