LaunchDarkly made the Inc. 5000 list 🎉
We’re one of the fastest-growing private companies in America.
Big thanks to our team and customers, we’re just getting started.
🚀 www.inc.com/inc5000
Hot take from our SVP of Product: building AI features ≠ managing dynamic learning systems.
Jenn Wei breaks it down on the What the Dev? podcast.
Culture, leadership, Experimentation, it all matters.
🎧 Ep 320: bit.ly/40RWa1h
A feature flag might’ve stopped the Google Cloud outage.
On June 12, a bad release hit global load balancers...no flag, no rollback.
Even the best teams need control after deployment.
Deployment isn’t the finish line.
Multiple environments. Multiple regions. One mistake? Global mess.
Guarded Rollouts give you the control to deploy safely across prod instances—US, EU, APAC, and beyond, without losing sleep.
Tame the chaos: bit.ly/4lEsJHI
Deploying before the weekend? Bold move.
But with Guarded Releases:
– issues get flagged automatically
– rollbacks happen instantly
– and you’re not stuck firefighting on Saturday
Smarter safety nets → smarter shipping.
Is your GenAI app hallucinating? Your users notice.
Fix it with runtime control:
→RAG with observability
→ LLM-as-judge scoring
→ Real time rollback with AI Configs
Stop relying on luck, start controlling your output: bit.ly/44orSEc
Create a flag. Roll it out. Ship it safely.
All from your editor.
LaunchDarkly’s new MCP server makes Cursor even more powerful—no context switching, no redeploys.
Tired of fragmented insights and delayed feedback in your feature development? 😩 LaunchDarkly Head of Analytics Divyansh Saini dives into why legacy approaches fail and how the shift to warehouse-native analytics is revolutionizing continuous improvement cycles.
Stop redeploying to tweak prompts.
With AI Configs, you can ship, experiment, and optimize your GenAI features, all at runtime.
Control your AI releases → bit.ly/4kPjVyL
Feature shipped. Now what?
Run warehouse-native experiments, unlock insights with Product Analytics, and iterate fast.
Smarter testing starts here → bit.ly/4n7RIoz
Your release shouldn’t feel like a gamble.
Guarded Releases provide real time visibility, instant rollbacks, and built-in safeguards, so you can ship with confidence, not chaos.
Get started → bit.ly/4n399GG
Calling all vibe coders!
@cursor.com.web.brid.gy is powerful out of the box, but way better with a few tweaks.
These 4 quick hacks will make your AI agent more accurate, personalized, and downright faster.
→ launchdarkly.com/docs/tutoria...
As a security measure, you’ll need to click the Run Tool button before the AI Agent will actually make any API calls. After the API calls have completed, switch to the LaunchDarkly UI. You can see that the flag has been created!
Now it’s time to create a flag! 🏁 Ask the AI agent to create a flag using the following prompt. Replace example-project with your LaunchDarkly project. “Create a feature flag in example-project to gate my awesome new feature.”
In @cursor.com.web.brid.gy, go to Settings > Cursor Settings > MCP > Add a new MCP server. Copy the following JSON, replacing the dummy key in this example with your API token.
To use the LaunchDarkly MCP server, you’ll need to create a LaunchDarkly personal API access token with Write permissions. You can do so on your Account Authorization page. Copy the token somewhere safe, since you won’t be able to view it again.
This thread will guide you through the process of installing the LaunchDarkly MCP server in the Cursor IDE. However, you can use our server with any AI client that supports MCP, such as Claude, @windsurfai.bsky.social AI, or even @raycast.com!
MCP, or Model-Context Protocol, is an open protocol that enables interaction with APIs using natural language. LaunchDarkly’s MCP server is powered by Speakeasy, a developer tool for generating SDKs, documentation, and agent tools.
You can now manage feature flags in plain English from the comfort of your IDE, using the LaunchDarkly platform, feature flags, and the new MCP server. Let’s take a look and learn how
AI regulation is fragmented and evolving fast. One-size-fits-all delivery won’t cut it.
Here’s how region-specific models and progressive rollouts help teams ship responsibly and stay compliant.
Read the post: aijourn.com/the-patchwor...
Want to see how feature flags power dynamic experiences behind the scenes?
See how a developer used #LaunchDarkly + Flask to toggle between jokes and data with one flag.
Learn how to do it yourself: bit.ly/3Gd0gcI
What do Easter eggs and feature flags have in common?
They both make your app more fun and easier to manage.
See how one developer used LaunchDarkly + Flask to toggle jokes and data with a simple flag.
🐣 Read the blog: bit.ly/3Gd0gcI
Building AI apps is exciting. Shipping them is another story.
Our new guide teaches how to manage unpredictable models, iterate faster, and scale safely.
Download Ebook: bit.ly/3EgGBIw
Hop into the full tutorial for more details 🐇
Crack Some Jokes With Feature Flags This Easter: bit.ly/3Gd0gcI
Happy coding and Happy Easter! 🐣🌷
Surprise your friends with more than just hidden Easter eggs! 🪺
Head to the LaunchDarkly dashboard and toggle the switch and crack a joke. 🍳
Write the Python script that checks if the environment key is set correctly to initialize the #LaunchDarkly SDK. 🥕🐇
Copy-friendly code is available at: bit.ly/3Gd0gcI
Feature Flag name: “change data source”
Configuration: Custom
Description: when enabled, changes the data source from egg facts to egg jokes.
Feature Flag type: String
Variants: tell a joke
Variants: share a fact
Serve when targeting is ON: tell a joke
Serve when targeting is OFF: share a fact
Build a basic Flask application that evaluates the flag in the Python Flask application and determines which data source to display on the website. 🌼 🧺 🌿
Copy-friendly code is available at: bit.ly/3Gd0gcI
🐣 This Easter, let's add humor to your Flask app and hide jokes behind an informative web page!
We'll use feature flags to toggle between egg-citing jokes and informative messages.
Ready to crack some code? 🥚😄
