Hey Product Hunt 👋 I'm Tidiane, founder of SafeRun. Every team I've talked to running AI agents in production has the same story. Their agent did something nobody could explain. The damage was already done by the time they noticed. The tools they had only logged the failure after the fact. One engineer told me he spent an entire weekend rerunning an agent trying to reproduce one failure. Another watched his sales agent email the same lead twelve times in five minutes. A third issued a $4,500 refund because the customer asked nicely. These aren't edge cases. This is what production AI agents do when they're given real tools and real money — and the current generation of monitoring tools tells you about it after the fact. I'm building SafeRun to close that gap. SafeRun sits inline between AI agents and the tools they call. It does four things: → Validates every tool call against your policies before execution → Breaks runaway loops and circuit-breaks on cost overruns → Escalates ambiguous actions to a human approval queue → Replays every agent decision frame by frame when something breaks The killer feature, based on what engineers keep telling me, is Replay. Step through every input, model thought, tool argument, policy result, latency, and cost — for every decision the agent made. Rerun from any step with modified inputs. Like a flight recorder for AI agents. We're integration-first: Python and TypeScript SDKs, with native support for LangGraph, OpenAI Agents SDK, Anthropic Claude Agent SDK, Vercel AI SDK, CrewAI, and Mastra. Or sit at the MCP layer for framework-agnostic coverage. Three lines of code to integrate. Pricing starts free for solo developers (1 agent, 10K actions/month, forever), then scales per-agent for teams. Two asks for the PH community: 1. If you're shipping AI agents to production, I'd love to have you in the early access batch — link is at the top of this page. 2. Tell me your worst AI agent failure story. Drop it in the comments below. The weirder, the better. I'm collecting them — the pattern across these stories is what shapes what gets built first. Massive thanks to everyone who supported the launch and to the early users who shaped what SafeRun became. And to whoever's reading this in the middle of debugging an agent at 2am — you're exactly who this is for. — Tidiane Founder, SafeRun saferun.dev

SafeRun

Reliability infrastructure for AI agents in production.

Reliability infrastructure for AI agents in production.