Introducing Release Autopilot: Automated release control for AI-built code — Read the announcement

Skip to:

Pricing for every stage.

  • Developer

    One platform for features, agents, observability, and experiments.

    Get started free
    $0per mo, foreverNo credit card required
    Features include:
    • Unlimited seats, no per-seat feesNo per-seat pricing. Invite your whole team; pricing is based on usage.
    • Unlimited flags with targetingCreate as many feature flags as you need and target them by user, segment, or rule.
    • Agent configs & LLM-as-judge evalsConfigure, evaluate, roll out, and observe your AI end to end.
    • Observability for apps & LLMsSession replays, errors, logs, and traces for your apps and the LLM calls inside them.
    • A/B tests & multi-armed banditsNo limit on experimentation.
    NewRelease AutopilotRelease AutopilotNewYour agents write the code. Autopilot ships it safely, and rolls it back if it breaks.See all capabilities
    Usage limits and rates
    Usage included, free every month
    Release AutopilotRelease AutopilotNewYour agents write the code. Autopilot ships it safely, and rolls it back if it breaks.See all capabilitiesIncluded
    Service connectionsOne server-side SDK connected to one environment, counted by time and averaged over the month. Every replica, worker and dev or staging environment counts.5 free
    Client-side MAUEach unique client-side user or device counts as one monthly active user.1K free
    AgentControl AI runsOne AI run is one LLM call your app makes through an AgentControl config, or one judge evaluation. Tokens and traces don’t count. An online eval scores a live response, so it uses two runs.5K free
    Session replaysRecordings of real user visits to your web or mobile app, captured by the LaunchDarkly observability SDK.5K free
    Errors capturedExceptions and crashes thrown by your own application code and sent to LaunchDarkly through the observability SDK or OpenTelemetry.5K free
    Trace spansDistributed traces from your services, sent through OpenTelemetry.10M free
    Log linesLog records your application sends to LaunchDarkly through the observability SDK or an OpenTelemetry collector, so you can read them next to flag changes.10M free
    Never billed. Upgrade when you hit a limit.

  • Foundation

    Ship features and agents to production. Scale both without hitting a ceiling.

    Pay as you go
    Pay as you go
    Everything in Developer, plus:
    • Flag & experiment approval requestsRequest a review before a change applies, with 1 to 5 reviewers. Making approvals mandatory requires Enterprise.
    • Unlimited projects & environmentsOrganize flags across as many projects and environments as you need.
    • Pay as you go past the free limitsKeep going past the free amounts and pay only for the usage you add.
    • Discounted annual billingSave 17% on usage rates when you bill yearly instead of monthly.
    • AWS/Vercel Marketplace billing availablePay for LaunchDarkly through your AWS Marketplace or Vercel Marketplace account.
    NewRelease AutopilotRelease AutopilotNewYour agents write the code. Autopilot ships it safely, and rolls it back if it breaks.See all capabilities
    Usage limits and rates
    Usage included, then pay as you go
    Release AutopilotRelease AutopilotNewYour agents write the code. Autopilot ships it safely, and rolls it back if it breaks.See all capabilitiesFree until Feb 1, 2027
    Service connectionsOne server-side SDK connected to one environment, counted by time and averaged over the month. Every replica, worker and dev or staging environment counts.5 free, then $10 each billed yearly ($12 each billed monthly)
    Client-side MAUEach unique client-side user or device counts as one monthly active user.1K free, then $8.33/1K billed yearly ($10/1K billed monthly)
    AgentControl AI runsOne AI run is one LLM call your app makes through an AgentControl config, or one judge evaluation. Tokens and traces don’t count. An online eval scores a live response, so it uses two runs.5K free, then $5/1K
    Session replaysRecordings of real user visits to your web or mobile app, captured by the LaunchDarkly observability SDK.5K free, then $2.92/1K billed yearly ($3.50/1K billed monthly)
    Errors capturedExceptions and crashes thrown by your own application code and sent to LaunchDarkly through the observability SDK or OpenTelemetry.5K free, then $0.25/1K billed yearly ($0.30/1K billed monthly)
    Trace spansDistributed traces from your services, sent through OpenTelemetry.10M free, then $1.25/1M billed yearly ($1.50/1M billed monthly)
    Log linesLog records your application sends to LaunchDarkly through the observability SDK or an OpenTelemetry collector, so you can read them next to flag changes.10M free, then $1.25/1M billed yearly ($1.50/1M billed monthly)
    Example: 10 connections and 2K MAU = $58.33/mo billed yearly ($70/mo billed monthly). 5 connections over the free 5 at $10 each, plus 1K MAU over the free 1K at $8.33 per 1K, billed yearly.

  • Enterprise

    Full platform controls and security for engineering teams at scale.

    Get custom pricingGet Enterprise trial
    Custom pricing
    Everything in Foundation, plus:
    • SCIM provisioning & IdP team syncProvision, update, and deprovision members, and sync teams, from your identity provider.
    • Custom roles & unlimited audit historyWrite your own roles with resource-level policies, group members into teams, and keep an unlimited history of who changed what.
    • Required approvals, scheduling, & workflowsMake approvals mandatory per environment or by tag, add teams as reviewers, route through ServiceNow or custom approval integrations, schedule changes, and build release workflows.
    • Bulk targeting & code referencesBulk targeting plus large list-based and synced segments, and see where each flag lives in your code.
    • CSM & premium supportA named customer success manager, plus Silver support with Gold or Platinum available.
    NewRelease AutopilotRelease AutopilotNewYour agents write the code. Autopilot ships it safely, and rolls it back if it breaks.See all capabilities
    Usage limits and rates
    Volume discounts, sized in your contract
    Release AutopilotRelease AutopilotNewYour agents write the code. Autopilot ships it safely, and rolls it back if it breaks.See all capabilitiesCustom
    Service connectionsOne server-side SDK connected to one environment, counted by time and averaged over the month. Every replica, worker and dev or staging environment counts.Custom
    Client-side MAUEach unique client-side user or device counts as one monthly active user.Custom
    AgentControl AI runsOne AI run is one LLM call your app makes through an AgentControl config, or one judge evaluation. Tokens and traces don’t count. An online eval scores a live response, so it uses two runs.Custom
    Session replaysRecordings of real user visits to your web or mobile app, captured by the LaunchDarkly observability SDK.Custom
    Errors capturedExceptions and crashes thrown by your own application code and sent to LaunchDarkly through the observability SDK or OpenTelemetry.Custom
    Trace spansDistributed traces from your services, sent through OpenTelemetry.Custom
    Log linesLog records your application sends to LaunchDarkly through the observability SDK or an OpenTelemetry collector, so you can read them next to flag changes.Custom
    Includes architecture mapping, security review, & POCs.. Before your contract is sized, our team maps your architecture, reviews your security needs, and runs a proof of concept with you.

Compare all features

Developer

Get started freeStart free

Foundation

Pay as you goPay as you go

Enterprise

Get custom pricingGet custom pricing

Platform

Platform features by plan

FeatureDeveloperFoundationEnterprise
AI Runs / month5000 / mo5000 / mo + overagesCustom (contracted)
Data / trace retention14 days30 days100+ days
Config delivery latency (SLO)<300ms<200ms<200ms
Overage behaviorIngestion stops at 5K$5 / 1000Contracted true-up
Add-on AI Run blocksNot included$5 / 1000Volume discounts available
Service Connections (SCs)Limited (free tier)SC-based billingSC-based billing
Client-side MAULimited (free tier)$10 / mo per 1K MAUCustom
Projects & environments1 project & 3 environmentsUnlimitedUnlimited

Agent configuration features by plan

FeatureDeveloperFoundationEnterprise
Prompts & model configsIncludedIncludedIncluded
Prompt SnippetsIncludedIncludedIncluded
Tools LibraryIncludedIncludedIncluded
Agent GraphsIncludedIncludedIncluded

Feature management features by plan

FeatureDeveloperFoundationEnterprise
Boolean feature flagsUnlimitedUnlimitedUnlimited
Multivariate flagsUnlimitedUnlimitedUnlimited
User / account / device targetingIncludedIncludedIncluded
Segments & percentage rolloutsIncludedIncludedIncluded
Flag historyIncludedIncludedIncluded
Code referencesNot includedNot includedIncluded
Flag schedulingNot includedNot includedIncluded
Workflows & approval (flag-level)Not includedNot includedIncluded

Evaluation & quality features by plan

FeatureDeveloperFoundationEnterprise
PlaygroundIncludedIncludedIncluded
Offline Evals (Datasets + Batch)IncludedIncludedIncluded
Online Evals (LLM-as-judge)IncludedIncludedIncluded
Custom JudgesIncludedIncludedIncluded
DatasetsIncludedIncludedIncluded

Experimentation features by plan

FeatureDeveloperFoundationEnterprise
A/B/n testing (configs)IncludedIncludedIncluded
A/B/n testing (feature flags)IncludedIncludedIncluded
Full-stack experimentationIncludedIncludedIncluded
Bayesian and frequentist analysisIncludedIncludedIncluded
CUPED variance reductionIncludedIncludedIncluded
Sequential testingIncludedIncludedIncluded
Sample ratio mismatch (SRM) detectionIncludedIncludedIncluded
A/A validity testingIncludedIncludedIncluded
Multi-armed banditsIncludedIncludedIncluded
Confidence and credible interval reportingIncludedIncludedIncluded
Flexible metric design: click, pageview, numeric, conversion, custom event, groupsIncludedIncludedIncluded
Warehouse-native integration (Snowflake, Redshift, Databricks, BigQuery and more)IncludedIncludedIncluded
Metrics import: API, integrations, SDK-levelIncludedIncludedIncluded
Metric measurement windowIncludedIncludedIncluded
Custom SQL metric definitionsIncludedIncludedIncluded
Mutual exclusion and layered experimentsIncludedIncludedIncluded
HoldoutsIncludedIncludedIncluded
Audience definitions and segment targetingIncludedIncludedIncluded
Results filtering by audience attributesIncludedIncludedIncluded
Experiment lifecycle stagesIncludedIncludedIncluded
Experiment ApprovalsNot includedNot includedIncluded

Observability features by plan

FeatureDeveloperFoundationEnterprise
AI Insights dashboardIncludedIncludedIncluded
LLM Observability (traces)IncludedIncludedIncluded
Data export (AI traces & metrics)Not includedIncludedCustom
Session Replays5K / moScalableCustom
Application Logs10M / moScalableCustom
Application Traces10M / moScalableCustom
Error monitoringLimitedScalableCustom

Release management features by plan

FeatureDeveloperFoundationEnterprise
Release automation
LaunchDarkly GitHub app: Reviews every PR, recommends a flag and a release plan, and applies both after one approval. Acts as the PR author and can be a required check before merge.Included, Free (limited)Included, Free until Feb 2027Included
Change classifier (Beta): Decides whether a change needs a flag, with a risk score, from a PR or your coding agent.Included, Free (limited)Included, Free until Feb 2027Included
Auto-flagging (Beta): Wraps a change in a feature flag and creates the flag in LaunchDarkly, from your coding agent or the PR. Undo with one command, and bring your own model key.Included, Free (limited)Included, Free until Feb 2027Included
Auto-release: Takes a flag from merge to fully released in every environment, following deploy signals and its release plan. Available through the API, CLI and MCP.Included, Free (limited)Included, Free until Feb 2027Included
Release plans: Define your release path once (environments, deploy signals, approvals) and reuse it across flags.Included, Free (limited)Included, Free until Feb 2027Included
Auto-cleanup: Checks that a flag is safe to remove, then opens a PR that deletes the dead code.Included, Free (limited)Included, Free until Feb 2027Included
Auto-archival (Beta): Archives finished flags automatically, after a grace period and a heads-up to the maintainer.Included, Free (limited)Included, Free until Feb 2027Included
Release agent: Chat with our agent to change how a PR is flagged or released.Included, Free (limited)Included, Free until Feb 2027Included
Release safety
Guarded releases: Roll out gradually while watching guardrail metrics, and roll back automatically on a regression. Set your own steps, with sample-ratio checks at each one, or choose notify-only.Included, Free (limited)Included, Free until Feb 2027Included
Multi-phase releases: Release to lower-risk audiences first, with a regression check before each phase moves on.Included, Free (limited)Included, Free until Feb 2027Included
Auto-fix: When a guarded release rolls back, see the errors it introduced and open a PR that fixes them.Included, Free (limited)Included, Free until Feb 2027Included
Adaptive triggers: Switch a flag or AI config to a safe variation automatically when a metric crosses its threshold.Included, Free (limited)Included, Free until Feb 2027Included
Flag audiences: See who got each variation of a flag, with their session replays. Filter by context, attribute, variation, date or release stage. Data is kept for 30 days.Included, Free (limited)Included, Free until Feb 2027Included
Release dashboard: See every active release in a project, its current stage and its health.Included, Free (limited)Included, Free until Feb 2027Included

Safety & guardrails features by plan

FeatureDeveloperFoundationEnterprise
AI Guardrails & BlocklistsBasicEnhancedAdvanced
PII RedactionNot includedNot includedIncluded

Integrations features by plan

FeatureDeveloperFoundationEnterprise
Model providersAll providersAll providersAll providers
Agent frameworksIncludedIncludedIncluded
API ControlsIncludedIncludedIncluded
Slack, Microsoft Teams, & TrelloIncludedIncludedIncluded

Governance features by plan

FeatureDeveloperFoundationEnterprise
SSO / SAML-- (2FA included)-- (2FA included)Included
SCIMNot includedNot includedIncluded
Audit LoggingNot includedNot includedIncluded
Advanced RBAC / Custom rolesNot includedNot includedIncluded
Approvals (governance-level)Not includedNot includedIncluded
Security & Compliance (SOC2 / ISO)Not includedNot includedIncluded

AI & agentic access features by plan

FeatureDeveloperFoundationEnterprise
MCP ServerIncludedIncludedIncluded

Support features by plan

FeatureDeveloperFoundationEnterprise
Support ResponseDiscord Community (1 business day outage)Discord Community (1 business day outage)1 business day (8 hours outage)
Support LevelBronzeBronzeSilver (Gold/Platinum Available)
Upgrade AvailableNot includedNot includedAvailable
Product UptimeNot includedNot includedGold/Platinum (99.99%)
Customer Success ManagersNot includedNot includedIncluded
Professional ServicesNot includedAvailableRecommended
Advisory SessionsNot includedAvailableIncluded
LaunchDarkly AcademyIncludedIncludedIncluded
Guided Onboarding ProgramNot includedAvailableIncluded
LaunchDarkly CertificationsIncludedIncludedIncluded

Resources to get started.