LaunchDarkly

What do you like best about the product?

I really like how easy and fast it is to pick up flag changes with LaunchDarkly. It's straightforward to use with Claude and MCP, which makes creating and configuring new flags a breeze. Plus, I can easily identify fully rolled out flags that can be removed, all with the help of Claude. The initial setup was easy too.

What do you dislike about the product?

I'd like it to proactively recommend flags I can remove, or identify misconfigured flags, like backend-only flags with client access or temporary flags that should become permanent.

What problems is the product solving and how is that benefiting you?

I use LaunchDarkly for safely rolling out changes to prod, experimenting, and swiftly reverting changes. It's fast and easy with Claude + MCP, letting me manage flags efficiently and identify unnecessary ones for removal.

What do you like best about the product?

Easy way to configure feature flags and configurations for multiple environments

What do you dislike about the product?

Interface that is hard to grasp, especially for the first timers

What problems is the product solving and how is that benefiting you?

It helps me to easily setup feature flags, target different users, run A/B tests

What do you like best about the product?

It's easy to integrate, reliable, fast, and relatively good UI to visualize the feature flags. The CLI is very useful to pilot FF through ai agents.

What do you dislike about the product?

Launch Darkly offers a lot of features, which is great, but if you’re only using a subset of them, the pricing can feel relatively very expensive.

What problems is the product solving and how is that benefiting you?

It helps us beta-test new features with a small segment of our audience, so we can gather feedback before opening them up to a larger audience.

What do you like best about the product?

As we develop our insurance claims categorization system for adjusters, doing full production releases frequently caused issues. LaunchDarkly's percentage-based progressive rollouts solved this by allowing me to roll out a new prompt or model update to just 5% or 10% of live claim extraction jobs rather than 100% at once. This is incredibly valuable. I also appreciate the in-memory flag evaluation, which ensures zero added network latency to our backend services because the Python SDK evaluates flag rules locally instead of making HTTP requests. The emergency kill switch is a major feature for me too; if an API starts failing, I can toggle the dashboard to instantly revert traffic back to our fallback logic without redeploying code. This makes handling issues seamless.

What do you dislike about the product?

There are a few things that could be improved in LaunchDarkly. Managing the lifecycle of feature flags requires strict engineering discipline. If developers don't clean up and archive old flags after a feature is fully rolled out, the dashboard becomes a confusing mess of stale toggles.

What problems is the product solving and how is that benefiting you?

LaunchDarkly decouples deployment from feature release, letting us safely test and implement experimental logic by toggling features internally. It reduces operational risks by allowing quick model switches, minimizing downtime without redeployment. Progressive rollouts and kill switches make release management smoother and more flexible.

What do you like best about the product?

I find LaunchDarkly to be simple and easy to use, which is great. I also appreciate that it's very lightweight. The feature that allows for managing different environments to turn features on and off is particularly useful. Overall, I liked pretty much all aspects of it.

What do you dislike about the product?

Sometimes can be overwhelmed

What problems is the product solving and how is that benefiting you?

LaunchDarkly helps me with testing and deployment of features using feature flags across different environments, and I find it simple, easy, and lightweight.

What do you like best about the product?

I like how LaunchDarkly makes feature releases less risky. We can roll out features gradually, test with specific users and quickly turn them off if something goes wrong without needing another deployment.

What do you dislike about the product?

It can feel a bit overwhelming at first, especially with a lot of feature flags across different environments. Keeping old flags cleaned up also takes some effort.

What problems is the product solving and how is that benefiting you?

LaunchDarkly helps us release features safely without tying every change to a deployment. We can test with a small group, roll out gradually, and disable features quickly if issues come up.

What do you like best about the product?

I really appreciate LaunchDarkly's intuitive user interface, which makes it easy for me as a product manager to quickly perform my tasks. The platform's ability to let me control feature flagging without needing to rely on engineering teams is a big plus. I can efficiently manage feature rollouts, create experiment cohorts, and have the flexibility to control who gets to access specific features. The onboarding process for me was straightforward and easy to understand, allowing me to get up to speed smoothly. It's a great feature launch platform for me.

What do you dislike about the product?

Creating segments in LaunchDarkly is a bit problematic for me. I have to add each customer manually to a segment one by one, which can be quite cumbersome. I would appreciate a more intuitive method for creating customer cohorts, rather than relying on CSV uploads or manual additions. This aspect of the platform could definitely be improved.

What problems is the product solving and how is that benefiting you?

I use LaunchDarkly to manage feature launches by experimenting with cohorts, quickly rolling back features if necessary, and handling feature flags without engineering involvement. It simplifies tracking, control, and environment management.

What do you like best about the product?

I like LaunchDarkly for its ease of use, which makes setting up user segmentation for tracking app performance straightforward. It also simplifies cohort breaking as desired. The initial setup was simple, which made getting started quick and easy.

What do you dislike about the product?

I find it confusing to choose between similarly named features like "user - plan type" because there's no clear dictionary or guidance on how the data can be split based on the flag features we have. It would be helpful to have more clarity in this area.

What problems is the product solving and how is that benefiting you?

I find LaunchDarkly makes cohort splitting easier for tracking app performance, which is simple. It's easy to use, but could use a better dictionary for data splitting based on flag features, especially for similarly named features.

What do you like best about the product?

I don't like anything about LD. I manage it; another team uses it. And collectively it's not great.

What do you dislike about the product?

LaunchDarkly has abysmal support and absolutely no strategy for retaining historical data during migrations. The ability to mass manage isn't there. As a system admin, this website is not ideal for management, and I don't recommend it.

What problems is the product solving and how is that benefiting you?

As an admin, they've only created problems for me.

What do you like best about the product?

What I like best about LaunchDarkly is how much confidence it gives teams when releasing changes. Feature flags make it possible to ship code separately from releasing functionality, so we can merge work earlier, test in production-like conditions, and turn features on gradually for the right users or environments.

The targeting and rollout controls are especially valuable. Being able to enable a feature for internal users first, then a small percentage of customers, then broader segments, makes launches feel much less risky. If something behaves unexpectedly, the ability to toggle it off quickly without a redeploy is a huge operational safety net.

I also appreciate how useful it is beyond simple on/off flags. Experiments, user targeting, kill switches, and environment-specific flag management all fit into the same workflow, which helps product, engineering, and QA collaborate around releases instead of relying on ad hoc config changes or emergency hotfixes.

What do you dislike about the product?

The main downside is that LaunchDarkly can add operational overhead if teams are not disciplined about flag cleanup. It is easy for temporary release flags to become permanent clutter, and over time that can make the codebase harder to reason about.

What problems is the product solving and how is that benefiting you?

LaunchDarkly solves the problem of risky, all-or-nothing releases. Before using it, every production deployment carried more pressure because a feature going live often meant the code was live for everyone at once. Now we can deploy code behind feature flags, release gradually, and turn features on for internal teams, beta users, or small customer segments first.

That has improved both speed and reliability. We can merge and deploy work earlier without waiting for every feature to be launch-ready, which reduces release bottlenecks and lets QA, product, and engineering validate behavior in real environments. Percentage rollouts and targeting rules also make it easier to catch issues before they affect the full customer base.

It also gives us a practical safety net. If a feature causes unexpected behavior, we can disable it quickly without rolling back an entire deployment or pushing an emergency hotfix. That shortens incident response time from a code-release cycle to a configuration change, which is a major benefit for customer-facing systems.

Another problem it solves is coordination across teams. Product managers can manage launch timing, engineers can keep risky changes isolated, and support teams can understand which customers have access to which experience. Overall, LaunchDarkly helps us release more frequently, reduce deployment risk, and experiment with less operational overhead.