Show HN: AgentRun – open-source, Self-improving Managed Agents
Repository · https://github.com/parcha-ai/agentrun
First seen by Oriel on .
AI summary
AgentRun is an Apache-2.0 platform that adds learning, durable execution, workflows, approvals, spend controls, and a unified console to agent harnesses.
- Agents can retain notes, skills, and workflows in a git-backed learning folder, with code checks and grading before changes are committed.
- Runs persist in SQLite or Postgres, resume after server or runner failure, and avoid rerunning finished steps; interrupted writes are marked rather than retried.
- Typed JSON workflows support tools, model calls, decisions, branching, parallel mapping, bounded loops, nested workflows, and escalation to an agent.
- The platform supports pi, Claude Code, Codex, and OpenCode, plus sandbox, desktop, phone, and browser computers from multiple providers.
- It runs locally on macOS or Linux with Node 24.22.3 or newer, or deploys to a server; Docker enables shell-based agent sandboxes.
#llm agents #agent orchestration #durable execution #workflows
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hacker News: 133, latest observation ; +122 over observed period since
Comments
Comment count reported by the source.
- Hacker News: 37, latest observation ; +37 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hacker News | 50471650 | https://github.com/Parcha-ai/agentrun/blob/main/README.md Read discussion |