A broken third-party API shouldn't ruin your weekend. Diane walks through how to add a kill switch that lets you shut things down instantly—no redeploys, no drama. Graceful failure > firefighting.

Rolling out a feature to everyone at once? Bold move. Here’s a quick way to minimize blast radius with progressive delivery—so bugs only hit 5% of users (not all of them).

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:

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.

RT @clairevo: Friendly reminder: @LaunchDarkly does lots of work to keep your client side flags/experiments private.

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:

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.

RT @thenewstack: Tired of fragmented insights and delayed feedback in your feature development? 😩 @LaunchDarkly Head of Analytics @navydish….

LaunchDarkly just ranked #1 in @G2dotcom Summer 2025 Feature Management Grid! 💥. 98/100 satisfaction. Real reviews. Real teams. Real results. Thank you to all our customers 💙. Full report →

RT @thenewstack: Why software teams are merging feature delivery with real-time learning to build smarter products, faster. By @LaunchDark….

Stop redeploying to tweak prompts. With AI Configs, you can ship, experiment, and optimize your GenAI features, all at runtime. Control your AI releases →

Feature shipped. Now what?. Run warehouse-native experiments, unlock insights with Product Analytics, and iterate fast. Smarter testing starts here →

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 →

Calling all vibe coders!. @cursor_ai 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. →

Honored to be named to the 2025 InfraRed 100 by @Redpoint!. LaunchDarkly helps teams build faster and smarter, driving the future of cloud infrastructure. See the full list:

Build, test, and ship—right where your data lives. LaunchDarkly + @Snowflake make Experimentation, product analytics, and AI config testing part of your release workflow. Smarter decisions, fewer silos, faster shipping.

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_ai, 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.