Openlit

Open sourceStar us on githubOpen Source Platform forAI EngineeringMonitor, debug, and improve your LLM applications with comprehensive observability, tracing, and evaluation tools. Built for production workloads.DocumentationView DemoSponsored byFeaturesPowerful Features for Modern TeamsEverything you need to build, ship, and scale your AI applicationsTracingEvaluationPrompt HubExperimentsDashboardsFleet HubDistributed TracingMonitor and trace your LLM applications in real-time. Visualize request flows, identify bottlenecks, and understand the complete lifecycle of every AI interaction with OpenTelemetry-powered distributed tracing.View DocumentationAI Model EvaluationRun online/offline evals via UI (experiment with prompts/models) and via SDKs (experiment with end-to-end application).View DocumentationPrompt ManagementCentrally manage, version, and deploy prompts across your applications. Experiment with different prompt variations, track performance, and iterate faster with version control for your prompts.View DocumentationExperiment with your prompts and modelsExperiment with your prompts and models to find the best performing ones. OpenGround is a playground for you to experiment with your prompts and models.View DocumentationReal-time MonitoringGet a unified view of all your LLM applications across different environments. Write custom SQL queries to analyze your AI telemetry data, create and resize custom widgets with flexible configurations and layouts and visualize telemetry from any OpenTelemetry-instrumented tool.View DocumentationMulti-Deployment ManagementGet a unified view of all your LLM applications across different environments. Monitor multiple deployments, compare performance metrics, and manage your entire AI fleet from a single dashboard.View DocumentationAll FeaturesDistributed TracingMonitor and trace your LLM applications in real-time. Visualize request flows, identify bottlenecks, and understand the complete lifecycle of every AI interaction with OpenTelemetry-powered distributed tracing.DocumentationAI Model EvaluationRun online/offline evals via UI (experiment with prompts/models) and via SDKs (experiment with end-to-end application).DocumentationPrompt ManagementCentrally manage, version, and deploy prompts across your applications. Experiment with different prompt variations, track performance, and iterate faster with version control for your prompts.DocumentationExperiment with your prompts and modelsExperiment with your prompts and models to find the best performing ones. OpenGround is a playground for you to experiment with your prompts and models.DocumentationReal-time MonitoringGet a unified view of all your LLM applications across different environments. Write custom SQL queries to analyze your AI telemetry data, create and resize custom widgets with flexible configurations and layouts and visualize telemetry from any OpenTelemetry-instrumented tool.DocumentationMulti-Deployment ManagementGet a unified view of all your LLM applications across different environments. Monitor multiple deployments, compare performance metrics, and manage your entire AI fleet from a single dashboard.DocumentationIntegrationGet started in minutesAdd comprehensive observability to your LLM applications with just a few lines of code. No code changes required for existing applications.Quick SetupGet OpenLit running in your environment in less than a minutePythonTypescriptmain.py1# Install OpenLit 2pip install openlit 3 4# Initialize in your application 5import openlit 6openlit.init()View Full DocumentationZero-Code Kubernetes ObservabilityAutomatically inject AI observability into your Kubernetes workloads without touching your codeAutomatic InstrumentationDeploy the OpenLIT Operator and it automatically instruments your AI applications - no code changes, no rebuildsOpenTelemetry NativeBuilt entirely on OpenTelemetry standards for seamless integration with your existing observability stackSimple ConfigurationJust create an AutoInstrumentation CR and select your workloads - that's it!Perfect for:LLM ApplicationsAI AgentsVector DatabasesAI FrameworksLearn About Kubernetes OperatorSupported IntegrationsWorks with all major LLM providers and frameworks out of the boxOpenaiLLMOllamaLLMAnthropicLLMDeepseekLLMGpt4allLLMCohereLLMMistralLLMGithub ModelsLLMVllmLLMAzure OpenaiLLMAzure Ai InferenceLLMHuggingfaceLLMAmazon BedrockLLMVertex AILLMGoogle Ai StudioLLMGroqLLMNvidia NimLLMXaiLLMElevenlabsLLMAi21LLMTogether.aiLLMAssembly AILLMFeatherlessLLMReka AILLMOLA KrutrimLLMTitan MLLLMSarvam AILLMPrem AILLMChromadbVector DBPineconeVector DBQdrantVector DBMilvusVector DBAstradbVector DBLangchainframeworkOpenai AgentsframeworkLitellmframeworkCrewaiframeworkLlama IndexframeworkBrowser UseframeworkPydanticframeworkDspyframeworkAg2frameworkHaystackframeworkMem0frameworkGuardrails AIframeworkPhidataframeworkMultionframeworkJulep AIframeworkLettaframeworkCrawl4aiframeworkFirecrawlframeworkdynamiqframeworkControlflowframeworkSwarmzeroframeworkNVIDIAgpusAMD RadeongpusWhy Choose OpenLit?Open Source & FreeAlways free, self-hosted, no vendor lock-inPrivacy FirstYour data never leaves your infrastructureProduction ReadyBuilt for scale with minimal performance overheadCommunityJoin our growing communityOpenLit is built by developers, for developers. Join thousands of engineers building better LLM applications with open-source observability.Open Source ProjectContribute to the future of LLM observability2323GitHub Stars5MSdk Downloads259Forks29Open IssuesView on GitHubJoin SlackSupport UsHelp us build the future of open-source LLM observabilityToken SupporterYou're literally counting tokens with us! Your support keeps our servers running, docs updated, and community thriving.❤️ Our eternal gratitude🌟 Help keep OpenLIT open-source🚀 Support faster developmentToken SupporterContext Window HeroYou've expanded our context window! Help us ship faster integrations, squash bugs, and build features the community loves.🏆 Your logo on our GitHub README🌐 Your logo on our official website🎯 Shape the future of AI observabilityContext Window HeroGet InvolvedWays to contribute to OpenLitContribute CodeHelp improve OpenLit with bug fixes and new featuresReport IssuesHelp us improve by reporting bugs and suggesting featuresWrite DocumentationHelp others get started with better docs and guidesShare Your StoryTell us how OpenLit helps your teamLatest from the BlogFleet Hub Playbook for Multi-Region AI ObservabilityNovember 7, 2025Monitoring LLM Usage in OpenWebUI with OpenLITFebruary 28, 2025How to protect your OpenAI/LLM Apps from Prompt Injection AttacksOctober 23, 2024View All PostsReady to Transform Your AI Observability?Join thousands of developers using OpenLIT to build better, more reliable LLM applications. Get started in less than a minute.Get Started FreeRead the Docs --- Pricing100% Open Source. Forever Free.OpenLIT is Apache 2.0 licensed. Self-host everything with zero licensing fees. A managed cloud option is coming soon for teams who prefer zero-ops.Available NowSelf-HostedRun on your own infrastructure$0/ foreverApache 2.0 — no license key, no usage limitsGet Started View on GitHubFull LLM observability (60+ integrations)OpenTelemetry-native traces & metricsToken usage & cost trackingGPU monitoring (NVIDIA + AMD)Vector DB monitoringPrompt Hub with versioningSecrets VaultFleet Hub (multi-deployment)LLM EvaluationsCustom model pricingManage organisationsExport to Grafana, Datadog, and any OTLP backendCommunity support (GitHub, Discord)Coming SoonCloudFully managed by the OpenLIT team—Join the waitlist to be notified at launchJoin WaitlistEverything in Self-Hosted (60+ integrations)Managed infrastructure — no ops overheadAutomatic upgradesDedicated supportSLA guaranteesSSO / SAMLAudit logsPriority feature requestsSupport the ProjectOpenLIT is free forever. If it saves you time or money, consider sponsoring development via OpenCollective.Token Supporter$10/monthShow your appreciation and help keep OpenLIT maintained.Sponsor on OpenCollectiveCommunity supporter badgePriority GitHub issue triageAccess to sponsor channel on DiscordContext Window Hero$50/monthDirectly fund new features and integrations.Sponsor on OpenCollectiveEverything in Token SupporterYour logo in README sponsors sectionDirect input on roadmap prioritiesMonthly call with core teamHow OpenLIT ComparesSee how OpenLIT stacks up against other LLM observability tools.OpenLIT vs LangfuseOpenLIT vs HeliconeOpenLIT vs LangSmithOpenLIT vs Datadog LLM Obs.Frequently Asked QuestionsIs OpenLIT really free?Yes. OpenLIT is Apache 2.0 licensed and free to self-host with no usage limits, no feature gates, and no license key required.What do I need to self-host?Docker and Docker Compose. Run `docker compose up -d` in the OpenLIT repo and you have the full stack running in under two minutes — UI, ClickHouse storage, and OpenTelemetry Collector included.Can I send data to my existing Grafana or Datadog instance?Yes. OpenLIT is OpenTelemetry-native. Configure the OTLP endpoint to point at any OTLP-compatible backend: Grafana, Datadog, New Relic, SigNoz, Jaeger, and more.Does OpenLIT support GPU monitoring?Yes. Enable GPU metrics with `openlit.init(collect_gpu_stats=True)`. OpenLIT collects utilization, VRAM usage, temperature, and power draw from NVIDIA and AMD GPUs.When will the Cloud tier be available?The OpenLIT Cloud managed service is in development. Join the waitlist by emailing contact@openlit.io and you will be notified at launch.Ready to Transform Your AI Observability?Join thousands of developers using OpenLIT to build better, more reliable LLM applications. Get started in less than a minute.Get Started FreeRead the Docs --- About OpenLITBuilding the Future of AI ObservabilityOpenLIT is an open-source AI engineering platform that helps teams build, evaluate, and observe AI applications across the entire lifecycle from development to production.Our MissionOpenTelemetry-native LLM Application ObservabilityOpenLIT is designed to make the integration of observability into GenAI projects as easy as pie – literally, with just a single line of code. Whether you're working with popular LLM libraries such as OpenAI and HuggingFace or leveraging vector databases like ChromaDB, OpenLIT ensures your applications are monitored seamlessly.We proudly follow the Semantic Conventions of the OpenTelemetry community, consistently updating to align with the latest standards in observability.View on GitHubDocumentationGrowing Open Source CommunityJoin thousands of developers building better LLM applications2323GitHub Stars5MSDK Downloads259Forks29Open IssuesWhat Makes OpenLIT SpecialA comprehensive platform with everything you need to build, monitor, and optimize your AI applicationsObservability & Real-time MonitoringDistributed tracing powered by OpenTelemetry with real-time insights. Track token usage, monitor response times, and debug AI applications in production.Fleet HubCentralized management and monitoring of multiple AI deployments. Get a unified view of all your LLM applications across environments.Prompt HubManage and version prompts for consistent and easy access across applications. Experiment and iterate faster.VaultSecurely handle your API keys and secrets centrally, avoiding insecure hardcoding practices.OpenGroundExplore, test, and compare various LLMs side by side. Find the best model for your use case.Kubernetes OperatorZero-code AI observability for Kubernetes workloads. Automatic instrumentation without touching your code.What is LIT?LIT stands for Learning Interpretability Tool. It refers to a visual, interactive model-understanding and data visualization tool, a term introduced by Google. OpenLIT builds on this concept to provide comprehensive observability for modern AI applications.Our Core ValuesOpen Source & FreeAlways free, self-hosted, and no vendor lock-in. Your data stays in your infrastructure.Privacy FirstYour sensitive AI data never leaves your infrastructure. Complete control and compliance.Community DrivenBuilt by developers, for developers. Every feature is shaped by community feedback.Get InvolvedWays to ContributeWhether it's big or small, we love contributions 💚. Here are a few ways to get involved:Contribute CodeHelp improve OpenLIT with bug fixes, new features, and integrations. Check our GitHub issues for good first issues.Report IssuesFound a bug? Have a feature request? Open an issue on GitHub and help us make OpenLIT better.Write DocumentationImprove our docs, write tutorials, or create examples. Help others get started with OpenLIT.Join the CommunityConnect with other users, share your experience, and help answer questions on Slack and Discord.Join Our CommunityConnect with the OpenLIT community and maintainers for support, discussions, and updatesGitHubStar, fork, and contributeSlackJoin live discussionsEmailContact us directlyX (Twitter)Follow for updatesSupport OpenLITHelp us build the future of open-source AI observability. Your support enables us to ship faster, add more integrations, and provide better support.Token SupporterContext Window HeroReady to Transform Your AI Observability?Join thousands of developers using OpenLIT to build better, more reliable LLM applications. Get started in less than a minute.Get Started FreeRead the Docs --- CompareOpenLIT vs the AlternativesSee how OpenLIT compares to other LLM observability and monitoring tools — feature by feature, with honest analysis.Why teams choose OpenLITOpenTelemetry-native — vendor-neutral, portable telemetry dataGPU monitoring for NVIDIA GPUs out of the boxSelf-hostable with Apache 2.0 license — free foreverVector DB monitoring alongside LLM tracing60+ integrations across LLMs, frameworks, Vector DBs, and GPUsPrompt Hub, Vault, and Fleet Hub includedDetailed ComparisonsOpenLIT vs LangfuseCompare OpenLIT and Langfuse for LLM observability. See how OpenTelemetry-native monitoring, GPU tracking, and self-hosted deployment stack up.See full comparison OpenLIT vs HeliconeCompare OpenLIT and Helicone for LLM observability. OpenLIT is SDK-based with no proxy latency, OpenTelemetry-native, and fully self-hostable.See full comparison OpenLIT vs LangSmithCompare OpenLIT and LangSmith for LLM monitoring. OpenLIT is open-source, self-hostable, and framework-agnostic. LangSmith is a proprietary cloud platform optimised for LangChain.See full comparison OpenLIT vs Datadog LLM ObservabilityCompare OpenLIT and Datadog for LLM observability. OpenLIT is purpose-built, open-source, and free to self-host. Datadog is a general-purpose platform with LLM monitoring as a paid add-on.See full comparison Ready to Transform Your AI Observability?Join thousands of developers using OpenLIT to build better, more reliable LLM applications. Get started in less than a minute.Get Started FreeRead the Docs