squick 1.3.0

Pre-computed LLM context for AI coding agents.

squick (Python)

Pre-computed, LLM-targeted code context for AI coding agents.

This PyPI package ships the right platform-specific binary inside the wheel and exposes it both as a squick console script and a small Python API. Installation is platform-aware: pip picks the wheel for your OS and architecture automatically.

Install

pip install squick

Use from the shell

squick scan ./your-project       # one-shot scan
squick watch ./your-project      # re-scan on file save
squick mcp                       # start an MCP server on stdio
python -m squick scan .          # equivalent to the above

Use from Python

import squick

# Run a scan and return the generated context.md as a string.
context = squick.scan("./your-project")
print(context)

# Locate the bundled binary, e.g. to spawn the MCP server manually.
print(squick.binary_path())

Supported platforms

OS	Architecture	Wheel tag
Linux	x86_64	manylinux2014_x86_64
Linux	aarch64	manylinux2014_aarch64
macOS	x86_64	macosx_10_12_x86_64
macOS	arm64	macosx_11_0_arm64
Windows	x86_64	win_amd64

See the main project README for full documentation, dictionary format, and MCP client configuration.

License

Apache-2.0. Copyright 2026 Horizon LLC. "Squick" is a trademark of Horizon LLC.