Show HN: AgentRun – open-source, Self-improving Managed Agents

8 relevance

Open-source managed agents platform with self-improvement directly matches AI agent orchestration interest and is highly actionable for developers.

AI/ML github.com

Answer the board on AI readiness

Arrive with evidence, not opinion. See how your foundation benchmarks against your peers, get a clear read, and leave with a ranked place to start.

Why It Matters

For platform engineers and AI/ML practitioners building agent systems, AgentRun directly addresses the exploding token costs and lack of state persistence that plague production agent deployments, offering a practical open-source alternative to expensive managed services.

Key Takeaways

Evaluate AgentRun as a drop-in control plane for your existing agent harness to reduce token spend and add self-improvement, starting with `npx @parcha/agentrun`.

AgentRun is an open-source platform that supercharges your agents. It adds self-improvement, durable execution and a control plane on top of your favorite agent harness. Your agents become smarter,...

Summary

AgentRun is an open-source platform that adds self-improvement, durable execution, and a control plane on top of agent harnesses like Claude Code, Codex, and OpenCode. It reduces token usage by persisting learned notes, skills, and workflows between runs, and provides fault-tolerant execution with run recovery, spend caps, and a unified console. A case study shows Greptile's code review costs dropped from $1 to $0.18 per review after tuning a cheap model on the user's own PRs, catching 1.6x more real bugs.