AgentRun
Open-source control plane for AI agents
Developer ToolsOpen sourceOpen source · verified#ai-agents#open-source#devops
Listed by the ProductPerch team from Hacker News · not launched here yet
What is AgentRun?
AgentRun is an open-source platform designed to manage and improve AI agents running on popular harnesses like Claude Code, Codex, or OpenCode. It provides a control plane that tracks runs, monitors spending, and handles approvals.
The system focuses on reducing token usage through self-improvement. After each task, the agent stores useful information such as notes, skills, or automated workflows. Subsequent runs start from this stored context instead of repeating the same steps.
Durable execution ensures that interrupted processes continue without data loss. If the server stops mid-run, restarting it resumes the work. The platform also supports external environments like sandboxes, Linux desktops, or Android devices. Users can deploy it locally with a single command using Node.js on macOS or Linux.
AgentRun features
- Self-improving memory that stores skills and workflows between runs
- Durable execution that resumes interrupted tasks after server restarts
- Centralized console for monitoring runs, approvals, and token costs
- Compatible with existing agent harnesses like Claude Code and Codex
- Supports external environments including sandboxes and Android devices
- Simple local deployment via npm with a single command
- Open source under the Apache-2.0 license
Who uses AgentRun and for what
- Automating code reviews while reducing model API costs
- Classifying merchant data by training cheap models on expert outputs
- Managing multiple AI agents across different development environments
- Running long-running research tasks that require uninterrupted execution
- Building self-hosted agent infrastructure for internal team workflows
AgentRun FAQ
Is AgentRun free?
Yes, AgentRun is released under the Apache-2.0 license and can be used at no cost. You only pay for the underlying AI model API keys you configure.
Is AgentRun open source?
Yes, the source code is publicly available on GitHub under the Apache-2.0 license.
What is AgentRun used for?
It manages AI agents by adding a control plane, durable execution, and self-improvement features to reduce token usage and track run history.
Does AgentRun work on Mac?
Yes, the platform requires Node.js and runs on macOS or Linux systems.
Can I use my own AI models with AgentRun?
Yes, you add your own model API key through the settings console to connect any supported provider.
How do I install AgentRun?
You can start it locally by running npx @parcha/agentrun, which launches the server and opens the browser console.
Comments (0)
Comments and upvotes open when this product launches on ProductPerch.
More Developer Tools launches
Monitor Laravel scheduled tasks and queues
Feel iOS haptic patterns on a real iPhone, copy the code
Deploy anywhere from your AI tool.
Python code quality analyzer for developers
Free product launch platform for indie makers
eBay listing data for apps and agents
