LaunchDarkly

What it is for

LaunchDarkly is a hosted feature-flag desk: you open Flags, and Developer is $0 / month, forever. Use it when the job is flags you toggle in production, not when you need a $20 chat seat.

What makes it unique: the cover is the official LaunchDarkly Flags list (example-json-flag Ready to archive, Henrietta Powell maintainer, Production Off / Test On, dates Nov 26, 2025) from the docs still — not a pricing card. The rows themselves are ops of Flags. LaunchDarkly bills service connections + client-side MAU, not a chat seat.

Key features

• Flags — a named toggle you turn On / Off per environment (cover still; docs). One list. Cover example-json-flag and Henrietta Powell sit here.
• Segments — a targeting group you reuse across flags (docs). A segment, not a flag. Rule reuse sits here.
• Experiments — an A/B test with a MAU meter (pricing 100K experimentation MAU). An experiment, not a flag. Guarded results sit here.
• Approvals — a change you queue before it ships (cover Approvals rail) (docs). An approval, not an experiment. Review sits here.
• AgentControl — an AI-run you observe and roll out (pricing 5k AI runs / month). An AI run, not a boolean flag. Eval sits here.

The SDK snippet is install. Developer / Foundation tiles are the meter. Session replay and error counts on the Foundation tile are add-on meters, not a sixth desk we fold into Flags.

Who it's for

Best fit: product / platform teams who will stay on Developer at $0 / month forever (unlimited seats, 5 service connections, 1K client-side MAU) while they open a first Flags list, then pay Foundation at $10 per service connection / month and $8.33 per 1k client-side MAU once usage is on the estimator.

Pricing

Verified 2026-10-23 (Asia/Shanghai) from launchdarkly.com/pricing. Tiles print Developer $0 / mo, forever and Foundation pay-as-you-go dollars.

• Developer — $0 / month, forever. Unlimited seats. 100K experimentation MAU / month. 10M logs and 10M traces / month. Unlimited feature flags. 5 service connections and 1K client-side MAU / month. 5K session replays and 5K errors / month. AgentControl: 5k AI runs / month, 14-day retention. No credit card.
• Foundation — pay as you go after a 14-day trial. $10 per service connection / month (billed yearly on the estimator we fetched). $8.33 per 1k client-side MAU / month. Extra AI runs $5 per additional 1k.
• Enterprise — custom pricing / Enterprise trial. Full-platform controls on the page.

Confirm checkout; taxes follow your location.

Limitations / costs

• Developer caps service connections at 5 and client-side MAU at 1K / month. A busier SDK fleet is the $10 Foundation connection.
• Paid dollars are usage (connections + MAU), not a $X/seat tile. Re-run the estimator before you budget.
• A flag desk, not an experiment-only product. For Statsig’s experiment desk see Statsig. For Flagsmith’s open-source flags see Flagsmith.

Alternatives

• Statsig — experiment-first flags when you want Statsig’s metric desk, not LaunchDarkly’s $0 Flags list.
• Flagsmith — open-source flags when you want Flagsmith’s self-host, not LaunchDarkly’s $10 service-connection meter.
• Unleash — open-source feature toggles when you want Unleash’s strategies, not LaunchDarkly’s AgentControl runs.

Official tutorial

Official LaunchDarkly channel video: Adaptive Triggers: Automatically Respond to Bad Releases (oEmbed author: LaunchDarkly). First-party release-response walkthrough; the Flags still is the cover. Adaptive Triggers on the video are a second desk, not generate-on-Flags.

Written path: launchdarkly.com/pricing plus the Flags still.

Website

https://launchdarkly.com

Repository

https://github.com/launchdarkly

Source

Catalog row curated for AI Tools Nav. Cover is the official LaunchDarkly Flags list from the docs CDN. Dollars checked on launchdarkly.com/pricing (Developer $0/mo forever; Foundation $10/service connection/mo and $8.33/1k MAU). Tutorial oEmbed author: LaunchDarkly.

Last verified

2026-10-23 (Asia/Shanghai) editorial check.