Mendral

14 billion log lines. That's how much CI data we processed across teams running the Mendral Agent in the last few weeks. One CI misconfiguration fired 654 times. A secret scanner not set up to handle direct pushes. Same error, same root cause. Some patterns worth sharing ↓

A package can be published to npm and pulled into your CI five minutes later. That's the attack window Dependabot-style scanners miss. Mendral's supply chain agent catches it at PR time. Typo-squatting, malware, suspicious post-install scripts, and packages under 72 hours old all get flagged before merge.

Custom agents are now available in Mendral. Build them in Slack: describe the task, iterate, turn it into an agent. Triggers: Schedule, GitHub Push/PR. A customer is already running daily CI failure triage, weekly build time reports, and CVE impact analysis on dependabot PRs.

New post: how we evaluate our CI diagnosis agent in production. 36,564 investigations. 5.7M CI jobs. 14.4B log lines. 60 days. No SWE-bench for CI. PR merge rate isn't accuracy. Confidence isn't calibrated. Two case studies of failures the agent caught itself. mendral.com/blog/how-we-kn…

Most CI dashboards are activity counters. Fixes merged. PRs reviewed. Numbers that look good and don't tell you much about what to do with them. We just shipped a redesign that flips it. Reactive panels for what's burning. Proactive panels for what's worth fixing before it burns. MTTR and P95 run duration as top-level KPIs. Mendral diagnoses CI failures, performance and security, and opens fix PRs autonomously. The dashboard is where you watch it work. Drop your email on mendral.com to get access.

Meet four more of the incredible speakers coming to AgentCon Silicon Valley. 🎤🔥 @SachaMorard @pamelafox @chrislatimer Sam Alba Founders, engineers, and advocates from Edgee, Microsoft, Vectorize, and Mendral — all bringing real-world experience to the stage at the Computer History Museum, Mountain View. Free to attend. Monday, 4 May 2026. Made possible by @pendoio @Microsoft @CircleCI @neo4j @descopeinc CODE Magazine @WeAreDevs 👉 Register now: agentcon.city/silicon-valley #AgentCon #GlobalAICommunity #AIAgents #AgentConSiliconValley #AIWorldTour

Mendral is at DevOpsDays Austin @DoDAustin, May 5-6. Catch us at the Startup Huddle for live demos and deep conversations on making CI fast and green: failures, flaky tests, slow feedback loops, and the people problems behind software delivery. Reply or DM @sam_alba to grab time. One spare ticket up for grabs. Let us know if you want it!

We're at AgentCon Silicon Valley on May 4th at the Computer History Museum in Mountain View. @sam_alba and @aluzzardi will talk about building a specialist AI agent for CI: why context stuffing doesn't scale, why we give the agent SQL access to ClickHouse, and how our multi-agent orchestrator handles over a billion log lines a week without blowing up cost or latency. If your CI is slow, noisy or flaky, come find us. globalai.community/chapters/san-f…

My full interview w/@sam_alba of @MendralHQ. Agents allow devs to generate 10x more code.Great for them, bad for ops.The pipeline (and ops) become overwhelmed. Mendrals answer, fight AI, w/AI, creating an agentic DevOps engineer that runs in the background youtube.com/watch?v=aZxQqF…

A key piece of our architecture, worth a read if you're building an agent 👇

Thanks @barton808 for the chat, it was a great opportunity to share what we’re doing.

AI is creating a SW delivery crisis and also giving us the tools to fix it - @sam_alba co-founder @MendralHQ (@ycombinator W26) talks about reinventing software delivery for the agent era, where he thinks the entire SDLC is heading and how Mendral can help youtu.be/aZxQqFiaDug

It was great chatting about the future of software, thank you @barton808.

When CI breaks, Mendral pulls logs, metrics, and alerts from your stack, diagnoses the root cause, and opens a fix PR. Shipped this week: infrastructure integrations. Connect @sentry, GCP, or @datadoghq. An AI DevOps engineer on duty, at every commit, for every repo.

Helping teams like PostHog, Metabase, Clipboard Health, and Luminai. You don't need their scale to have these problems → mendral.com

If someone fixes the issue outside Mendral, the agent detects it and auto-closes the insight. If the problem comes back, it reopens with full history. If two insights share a root cause, the agent merges them. After a few weeks, the agent knows every failure pattern in your CI.

14 billion log lines. That's how much CI data we processed across teams running the Mendral Agent in the last few weeks. One CI misconfiguration fired 654 times. A secret scanner not set up to handle direct pushes. Same error, same root cause. Some patterns worth sharing ↓