A walkthrough of Painted Wolf Code
Chris Beckman, creator of Painted Wolf Code, walks through the application in a new video demo.
In this walkthrough, Chris demonstrates how Painted Wolf Code works in practice: an open-source, local-first AI coding agent that runs on your Mac, connects directly to the model providers you already use with your own API keys, and enforces guardrails at the operating system level rather than relying on model restraint.
The demo covers:
- Local-first execution: Your code, transcripts, and session history remain on your machine, with native execution powered by Apple Silicon.
- Provider flexibility: Spreading work across frontier and fast models from multiple providers in a single session, with transparent cost tracking.
- Decision pruning with Bialy: How our built-in decision engine (built on Laya) sorts tools, skills, and code before frontier model calls, cutting prompt overhead and token costs.
- OS-enforced safety: Transparent approvals that only trigger when real risk is involved, backed by an extensive catalog of cloud and shell rules.
Getting started
Painted Wolf Code is open-source under the Apache-2.0 license and available now for macOS 14 (Sonoma) or later on Apple Silicon.
You can install it using Homebrew:
brew install --cask paintedwolf-ai/tap/painted-wolf-code
You can also download disk images from our Download page, review the full release notes on the Changelog, or inspect the source code on GitHub.