HawkeRun
← All posts
Press·1 min read

HawkeRun Launches a Local-First Supervisor That Keeps AI Agents Running

HawkeRun announces the launch of its local-first agent supervisor for Windows, macOS, and Linux, giving developers a durable runtime that keeps Claude, Codex, and other AI workers running unattended.

By Isabella Nevarez

FOR IMMEDIATE RELEASE

CodeHawke today announced the launch of HawkeRun, a local-first supervisor that keeps AI coding agents running while developers are away. HawkeRun gives developers a durable runtime that feeds tasks to Claude, Codex, Ollama, and other command-line workers, detects stalls, kills frozen processes, requeues work, and escalates only the decisions that truly need a human.

AI coding agents have become remarkably capable, but they are built around a person watching the screen. The moment that person steps away, agents stall on routine questions, hang on stuck subprocesses, and lose hours of progress. HawkeRun closes that gap.

"The supervisor is permanent, and the AI tools are disposable," said the CodeHawke team. "People should not have to babysit their agents. HawkeRun makes them work like a managed team that reviews itself, recovers from failure, and only knocks on your door when it matters."

Key capabilities

Availability

HawkeRun runs on Windows, macOS, and Linux. A free tier is available today, with Pro, Power, and Team plans for larger workflows. Developers can get started at hawkerun.com.

About CodeHawke

CodeHawke builds tooling that makes autonomous AI development reliable enough to trust unattended. For more information, visit hawkerun.com or contact support@hawkerun.com.