We did it again! 10 consecutive seasons at #1 in the G2 Grid Report for Feature Management. 98 overall score. 99 Satisfaction. 96 Market Presence. Basically, customers keep saying the same thing: when AI is writing more and more of what ships, they want LaunchDarkly holding the controls.
LaunchDarkly
Software Development
Oakland, CA 50,753 followers
Empowering all teams to deliver and control their software.
About us
LaunchDarkly is the runtime control layer for the AI era. For over a decade we have helped engineering teams control what ships in production, starting with feature management and evolving into the infrastructure that governs how software and AI agents behave after deploy. With CodeControl and AgentControl, teams can update agent behavior in milliseconds, automatically roll back when something drifts, and govern every agent and feature across the organization from one place. No redeploys required. LaunchDarkly processes more than 50 trillion flag evaluations per day and serves some of the most demanding engineering teams in the world, including over a quarter of the Fortune 500. Founded in 2014 in Oakland, California by Edith Harbaugh and John Kodumal, LaunchDarkly has been named to the Forbes Cloud 100 for five consecutive years, Fast Company's Most Innovative Companies list, and the Inc. 5000.
- Website
-
https://launchdarkly.com
External link for LaunchDarkly
- Industry
- Software Development
- Company size
- 501-1,000 employees
- Headquarters
- Oakland, CA
- Type
- Privately Held
- Founded
- 2014
- Specialties
- Continuous Delivery, Feature Flagging, DevOps, Feature Toggling, Feature Management, Experimentation, Technology Migration, Targeted User Experiences, and Mobile App Solutions
Products
![Click here to view LaunchDarkly]()
LaunchDarkly
DevOps Software
One platform to ship, govern, and continuously improve AI agents in the AI era. Engineering teams are shipping AI agents faster than ever. LaunchDarkly's AgentControl gives them the runtime control layer to match that pace, with real time visibility into production behavior, instant configuration changes without touching the codebase, and guardrails that catch problems before users feel them. When something drifts, Adaptive Triggers automatically correct it. When performance needs improving, Agent Optimization finds a better configuration for you.
Employees at LaunchDarkly
View 669 employees at LaunchDarkly
or
Locations
-
Primary
Get directions
1999 Harrison St
Suite 1100
Oakland, CA 94612, US
Updates
-
Somebody's going to demo an agent at The AI Conference that works perfectly right up until the Q&A. We've got a discount code for you in the comments 🫶 Grab time with us: https://bit.ly/4fJoARz
-
-
Jay Khatri's sitting on something that changes how you think about your alerts, and he's not saying what it does until the video drops. New episode of The Control Panel drops in two days!
-
We're interested in finding whoever's AI agent has already done something slightly unhinged in production and just hasn't mentioned it in standup yet. We'll be at Gartner APPS this September! 🎡💂🇬🇧 Grab time with us: https://bit.ly/468wsHQ
-
-
Two things already sitting in your LD account: 1. Catching breakage before your customer tells you. 2. Running experiments without ever exporting data out of your warehouse. Kellye King shows you exactly where to find both. New episode of The Control Panel drops August 26.
Watch live August 26 | 1 PM ET
-
We gave our agents a tool to complain and are calling it ✨vent✨ Turns out complaining is a great trigger for automation. One vent kicks off triage → Slack post so we can watch it happen live → Jira ticket → then a second agent decides whether to fix it itself or flag it for a human. Step 4: "reads the codebase and opens a PR."
-
-
Glad to see Cameron Etezadi's thinking on this is getting picked up. His point that intervention matters as much as deployment is one we'd fully stand behind.
What happens when an AI agent goes beyond what you expected and starts making decisions you never explicitly approved? As Agentic AI moves from experimentation into core enterprise operations, that question is becoming increasingly urgent. For engineering and technology leaders, the challenge is no longer simply deploying autonomous systems. It is maintaining visibility and control once those systems are live, adapting to new contexts, changing models and unexpected behaviour in real time. In this article for NODE, Cameron Etezadi, CTO at LaunchDarkly, explores why progressive rollouts, continuous observability and dynamic operational controls are becoming essential for organisations looking to scale Agentic AI safely. Because when software can act on our behalf, knowing how to intervene matters just as much as knowing how to deploy it. Read more here: https://lnkd.in/epq_mMW2 #AgenticAI #EnterpriseAI #AIGovernance #ArtificialIntelligence #SoftwareEngineering #TechLeadership #DigitalTransformation
-
-
We're heading to Delhi! Our Senior Solutions Engineer, Abhishek Hiremath, is taking the stage to talk about why deployment isn't the same as release anymore and what that shift means for teams moving beyond traditional CI/CD. → Catch his session: https://bit.ly/4hAAv6S
-
