🚨 GitHub just dropped their new AI coding agent for Pro & Enterprise users Naturally, I threw it straight into production to fix my real bugs. 😈🐞 It worked!! Here’s what happened 👇🧵

To reason, plan & fix in complex real-world systems, AI coding agents need more than just the code. They also need ubiquitous, function-level production context. Hud (Hud) captures exactly that, with zero configuration. Now agents know how code actually behaves 🤩

Great vibes at EuroPython today! 🔥 It was really exciting to discuss the future of monitoring in the era of #AI and showcase some of what we've built at Hud for self-healing production with coding agents like GitHub Copilot and Anthropic 's Claude CLI, utilizing Python Software Foundation

AWS just announced the new marketplace for AI Agents, and we are proud to be a launch partner! If you’re on Amazon Web Services it’s now much easier to get Hud - just search "fix production issues".. We’re honored to be in the first group included, alongside titans like Anthropic

Rolling out AI coding tools sounds simple: 1️⃣ Pick a vendor 2️⃣ Let your smartest devs try it 3️⃣ Watch productivity soar Reality? Most rollouts stall by week 3. At Drata , 200+ engineers use Augment weekly. Here’s the playbook they used (and what most teams miss) 🧵

At Hud we solve a core gap in engineering and agentic coding: generating code that actually works in complex production. The Runtime Code Sensor installs in under a minute. No config. No maintenance. Automatically captures live function-level data automatically. See what’s

What started as a coincidence (many team members wearing the same light purple Uniqlo shirt on a Thursday) became tradition and homage (on Thursdays we wear purple, #meangirlsforever), and like everything in startup lore - ended up becoming swag 💜 Hud

Just taking a moment of appreciation for the Hud founders committing to the bit

Geoffrey Litt Someone's already building a Hud for AI - that's us! - Hud - check out hud.io (especially the Hudders in the IDE - they were the hint that led to our name). Also - genius Top Secret reference! Happy to connect and share more about our approach...

Definitely agree on the runtime context part 👩‍💻Hud

Code isn’t just text - it’s alive. It behaves, sometimes in surprising ways. Hudders bring that reality into your IDE: live headers above each function showing error rates, usage, and more. Hear our co-founder Roee Adler explain all about them and check out hud.io!

With the right tools vibe bug fixing can be less of a nightmare MCP to bring observability data, instructions to use git to correlate to recent changes and linking it all correctly can make at least RCA a done problem Look at Hud for example, seems like they nailed it

“Even though observability is a $100 billion market, it still sucks and production is on fire and no one knows why.” clickhou.se/4qSMcrF May Walter and her team at Hud are using ClickHouse to transform observability with the world’s first runtime code sensor. See

We built a system that understands how code behaves in production: light when everything runs smoothly, forensic when it doesn’t. We needed a data engine that is super fast and built for high-cardinality at scale. That’s why we chose ClickHouse - read more in this case study:

AI isn’t showing up as magic anymore - it’s showing up in real production workflows. VentureBeat just covered how teams like monday.com and Drata boosted debugging speed by letting AI see runtime reality, not guesses. Agentic engineering is starting to look very real!

Roee Adler VentureBeat Hud monday.com Drata Congratulations! Your great team and technology absolutely deserve this spotlight. Really exciting to see this important recognition on VentureBeat. I’m sure it’s just the beginning! 😊🙏🏻

The Automations team at Monday.com made a deliberate choice to build a repeatable playbook for discovering, understanding, and quickly resolving complex performance issues as the platform evolves. Read about it here: hud.io/case-study/how…

We’ve been working with the Amazon Web Services team on integrating Hud into Kiro : install Hud in 1m, and have errors, performance degradations and CPU spikes detected with the root cause needed to fix them quickly. Give it a try!