Trace, evaluate, and improve AI agents with one open platform. Use production data to understand behavior, collaborate on fixes, and ship better quality at lower cost and latency.
Language: en
Indexed excerpt
Used by 21 of Fortune 5090B+ observations/month100,000+ engineers building on LangfuseUsed by 21 of Fortune 5090B+ observations/month100,000+ engineers building on LangfuseUsed by 21 of Fortune 5090B+ observations/month100,000+ engineers building on LangfuseOpen Source Agent Evals &ObservabilityTrace, evaluate, and improve AI agents with one open platform. Use production data to understand behavior, collaborate on fixes, and ship better quality at lower cost and latency.Start freeSDocumentationDOnboard with AIRead storyRead storyRead storyRead storyRead storyRead storyLaunch, observe, improve — repeat.Langfuse connects tracing, monitoring, datasets, experiments, and evaluation in one continuous loop. Use production signals to understand behavior, test improvements, and ship better agents with confidence.The full AI engineering loopSee how observability, prompts, evals, experiments, and human feedback work together.Learn in AcademyAll the tools, oneintegrated platform.One integrated platform to trace, manage prompts, evaluate, and experiment from prototype to production scale.ObservabilityHierarchical traces capture every LLM call, tool invocation, and retrieval step. Filter by user…
DocsAdministrationData DeletionCopy pageData Deletion There may be use-cases where you want to remove selected data from Langfuse, like erroneously created traces in a development flow, user data for PII, or your whole project. In case you want to retain only recent data, you can use our Data Retention feature. You can delete unwanted data from Langfuse by: Deleting a single trace; Deleting a batch of traces; Deleting all traces in a session through the API; Deleting all traces that match a query filter; Deleting a project; Deleting an organization; or Deleting a user account. Below, we will walk through each of the options and their guarantees. Deleting Traces Note that all trace deletions will delete related entities like scores and observations across all data storages. Single Trace Langfuse UIAPITo delete a single trace, open its detail view and hit the Delete button. Confirm that you want to delete the given trace.DELETE /api/public/traces/{traceId}See reference. Batch of Traces Langfuse UIAPITo delete a batch of traces, select them in the trace list and select Delete in the Actions dropdown.DELETE /api/public/tracesSee reference. Delete a session's traces through the API Whe…
A specialized decision model for routing, classification, and eval verdicts — and how to score Langfuse traces with it.
Language: en
Indexed excerpt
September 18, 2026Using TypeSafe's Jev for evalsA specialized decision model for routing, classification, and eval verdicts — and how to score Langfuse traces with it.Annabell Schäfer TypeSafe launched Jev mid-September 2026. It makes decisions, cannot write a single sentence, and is 20 to 200x faster and 40 to 400x cheaper than frontier models according to TypeSafe. You send a state, a string or JSON, plus typed questions. You get typed answers with probabilities. It gives you no reasoning back. Useless at other tasks. Decisions matter a lot for AI agents and AI-powered workflows. AI decides between different categories, whether something is true or not, which tool or path to take. And in the case of evals, AI classifies or simply decides if an output is right or wrong. Over the past years we forced this out of LLMs with structured outputs and JSON schemas, and still paid generation prices for a yes/no. Jev is deliberately built for that job instead of adapted to it. The flipside: it will not write code, summaries, or tell you why it answered the way it did. What Jev is good for Jev can help anywhere in your pipeline or eval harness where decisions are repeated, high volume, and t…
Who Langfuse is built for: engineering teams shipping AI applications that are core to their product, iterating constantly around model limitations.
Language: en
Indexed excerpt
HandbookChaptersWho are our customers?Copy pageWho are our customers? We are building Langfuse for engineering teams working on impactful AI applications/features that are core to their product strategy, constantly iterating and working around model limitations. We are working with 50,000+ companies using Langfuse across industries, AI use cases, company sizes and geographies. Some of our customers and what they do: Ramp - Smart financial infrastructure; self-hosted agent observability (Watch Story) Canva - Online design platform for creating visual content (Read Story) Intuit - Financial software company (TurboTax, QuickBooks, Mint) Juicebox.ai - AI recruiting platform that powers outbound talent sourcing Samsara - Fleet management and IoT platform for physical operations Twilio - Cloud communications platform (SMS, voice, video APIs) SumUp - Point-of-sale and payment processing solutions for small businesses (Read Story) Khan Academy - Free online educational platform with courses and lessons (Read Story) Magic Patterns - AI-powered design tool for generating UI components and prototypes (Read Story) Merck - Global pharmaceutical and life sciences company (Read Story) Hugging Fac…
With Langfuse you can capture all your LLM evaluations in one place. You can combine a variety of different evaluation metrics like model-based evaluations (LLM-as-a-Judge), human annotations or fully custom evaluation workflows via API/SDKs. This allows you to measure quality, tonality, factual accuracy, completeness, and other dimensions of your LLM application.
Language: en
Indexed excerpt
DocsEvaluationOverviewCopy pageEvaluation Overview Evals give you a repeatable check of your LLM application's behavior. You replace guesswork with data, and catch regressions before you ship a change. Evaluation runs across most of the AI engineering loop: you score live traces in production, turn interesting examples into datasets, run experiments to compare changes, and judge the results with manual or automated evaluators. It happens both online, on live production traces, and offline, before you ship a change. DeployOnlineTracetraces · sessions · agents · promptsOnlineMonitordashboards · LLM-as-judge · feedbackOfflineBuild datasetsdatasets · features-as-testsOfflineExperimentprompts · models · code variantsOfflineEvaluatejudges · custom evals · annotation 🚀Want to see it in action? Create a free account and explore Langfuse Evaluation in the interactive example project. Getting Started You can evaluate both: live incoming traces to measure quality on production data and track trends over time. your existing application on a pre-defined dataset, to make sure your changes are ready for production. For more information on how evaluators, scores, datasets, and experiments fit tog…
Try Langfuse in action with a live example project for free. Interact with the chatbot to see new traces and user feedback (👍/👎) in Langfuse. No credit card required.
Language: en
Indexed excerpt
DocsExample ProjectCopy pageExample Project The Langfuse example project is a live, shared project that lets you explore Langfuse's features with real data before setting up your own account. Step 01Generate your first demo traceRun a demo interaction below to generate real data.Step 02Find your trace in LangfuseOpen Tracing and select your latest request to inspect its trace.Step 03Explore the whole demo projectOpen the live shared project in Langfuse. 🎥Prefer video? Watch the end-to-end walkthrough of all Langfuse features. Generate your first demo trace Q&A ChatbotChatVoice AgentVoiceImage GeneratorImageSentiment ClassifierSentimentRock Paper ScissorsRPS👋 Do you have any questions about Langfuse?Ask me!What can I use Langfuse for?How do I link my prompts to my traces? My code is in pythonHow do I get started with tracing?⚠️ Warning: Do not enter sensitive information. All chat messages can be viewed in the public example project. Responses may be inaccurate. Please check the documentation for details or reach out to us via the chat widget.Click to start a voice conversationStore audio recording on the traceStart a voice conversation with the AI agent. It runs on OpenAI GPT-Liv…
Langfuse is an open-source AI engineering platform (GitHub) that helps teams collaboratively debug, analyze, and iterate on their AI agent applications. All platform features are natively integrated to accelerate the development workflow.
Language: en
Indexed excerpt
DocsOverviewCopy pageLangfuse Overview Langfuse is an open-source AI engineering platform (GitHub) that helps teams collaboratively debug, analyze, and iterate on their AI agent applications. All platform features are natively integrated to accelerate the development workflow. Langfuse is open, self-hostable, and extensible (why Langfuse?). ObservabilityLog tracesLowest level transparencyUnderstand cost and latencyPromptsVersion control and deployCollaborate on promptsTest prompts and modelsEvaluationMeasure output qualityMonitor production healthTest changes in developmentPlatformAPI-first architectureData exports to blob storageEnterprise security and administration Observability Observability is essential for understanding and debugging AI agent applications. Unlike traditional software, AI agent applications involve complex, non-deterministic interactions that can be challenging to monitor and debug. Langfuse provides comprehensive tracing capabilities that help you understand exactly what's happening in your application. Traces include all LLM and non-LLM calls, including retrieval, embedding, API calls, and more Support for tracking multi-turn conversations as sessions and us…
We've released new comprehensive walkthrough videos covering observability, prompt management, and evaluation to help you get up to speed quickly with Langfuse.
Langfuse August Update: New Observation Types, End-to-End Walkthrough Videos, Agent Demo Project, and more.
Language: en
Indexed excerpt
August 31, 2025Langfuse August UpdateLangfuse August Update: New Observation Types, End-to-End Walkthrough Videos, Agent Demo Project, and more.Marc Klingen This month, we've shipped updates to improve tracing for complex agents, enhance team collaboration, and expanded the integration library. Here's what's new: New Observation Types We've introduced new observation types to give you deeper insight into your agent's behavior. You can now semantically label spans as Agent, Tool, Chain, Retriever, Embedding, and Guardrail, making it easier to debug and analyze complex traces. For most agent framework integrations, these types are detected automatically. You can also set them manually via the Langfuse Tracing SDKs. → Learn more in the docs End-to-End Walkthrough Videos We've released many new capabilities over recent months. We recorded four new short videos covering observability, prompt management, and evaluation to help you get up to speed quickly. These have since been consolidated into a single end-to-end walkthrough. → Watch the walkthrough Agent Example Project Experiment with our revamped interactive example project (langfuse.com/demo) that uses the Langfuse Docs MCP server t…
Manage and version your prompts in Langfuse (open source). When retrieved, they are cached by the Langfuse SDKs for low latency.
Language: en
Indexed excerpt
DocsPrompt ManagementOverviewCopy pagePrompt Management Prompt management is a systematic approach to storing, versioning, and retrieving prompts for your LLM application. Instead of hardcoding prompts in your application code, you manage them centrally in Langfuse. 🚀Want to see it in action? Create a free account and explore Langfuse Prompt Management in the interactive example project. Decouple Prompt Updates from Code Deployment In most LLM applications, prompt iteration and code deployment are managed by different people. Product managers and domain experts iterate on prompts, while engineers manage deployments. With prompts in code, a simple text change requires engineering involvement, code review, and a full deployment cycle, turning a 2-minute update into hours or days of waiting. When prompts live in Langfuse, non-technical team members update them directly in the UI while your application automatically fetches the latest version. This separation of concerns means prompt updates deploy instantly, without needing to involve engineering or triggering a deployment. No latency, no availability risk Langfuse Prompt Management adds no latency to your application. Prompts are ca…
Open source application tracing and observability for LLM apps. Capture traces, monitor latency, track costs, and debug issues across OpenAI, LangChain, LlamaIndex, and more.
Language: en
Indexed excerpt
DocsObservabilityOverviewCopy pageObservability & Application Tracing Generative AI systems are inherently non-deterministic. Therefore, debugging your application without any observability tool is more like guesswork. Implementing comprehensive AI observability gives you the tools to understand what's happening inside your application and why. It is the foundation for continuously evaluating and improving your system. The core of this is application tracing — structured logs of every request that capture the exact prompt sent, the model's response, token usage, latency, and any tools or retrieval steps in between. Here's an example of a trace in the Langfuse UI: 🚀Want to see it in action? Create a free account and explore Langfuse Observability in the interactive example project. Getting Started Start by setting up your first trace, and compare it against our best practices guide so you're set up for success. If you're new to AI observability, take a look at the core concepts too. Use your traces Once traces are coming in, the next step is to make sense of them and use what you learn to improve your agent. Read the chapter on Monitoring in the Langfuse Academy to learn how to uti…
End-to-end examples and resources to get started with Langfuse for LLM Tracing, Monitoring, Prompt Management, and more.
Language: en
Indexed excerpt
GuidesGuidesCopy pageGuides End-to-end examples of how to use Langfuse: tutorials, runnable cookbooks, and video walkthroughs. Introduction to Langfuse If you are new to Langfuse, watch this walkthrough to get a thorough introduction to the Langfuse platform. Introduction to LangfuseGet an overview of the complete Langfuse platform and learn how it helps teams build better LLM applications through observability, prompt management, and evaluation.Technical documentation Evaluation Tutorials Start with the fundamentals and progress through advanced techniques as your application complexity grows. Read our blog post for an overview. Set Up a User Feedback LoopCapture explicit and implicit user feedback as scores on your traces, and use it to surface the responses worth reviewing.Error AnalysisReview traces to classify issues like hallucinations, irrelevance, and formatting errors. Turn raw logs into actionable insights.Automated EvaluationsAutomated evaluators give you a scalable way to measure and monitor your application’s failure modes.Testing LLM ApplicationsBuild a testing foundation with deterministic checks and probabilistic ones using LLM judges to prevent regressions.CI/CD fo…
Langfuse is the open-source Braintrust alternative. September 2026 comparison of MIT self-hosting, ClickHouse vs Brainstore, unit vs GB pricing, and evals.
Language: en
Indexed excerpt
CompareBraintrustCopy pageLangfuse vs. Braintrust This guide outlines the key differences between Langfuse and Braintrust to help engineering teams choose the right AI engineering platform. All Braintrust facts were checked against public Braintrust sources in September 2026. TL;DR Choose Langfuse if you want: 1. Open source and full control. MIT-licensed self-hosting of the entire application on every tier, with data in a database you can query directly. 2. Production observability at scale. Complete traces at high volume on ClickHouse, open source since 2016, with dashboards, alerts, and full-text search. Observability is the foundation; evals, prompts, and datasets layer on the same data. 3. Predictable, count-based pricing. A trace, observation, or score is one unit: cost follows request count, not payload size, and scores are not a separate meter. Pro includes 3-year retention. Choose Braintrust if your team is evaluation-first and wants a vendor-managed workbench for reviewing experiments, with AI-assisted analysis built in. Open source and distribution The core difference is the distribution model. Langfuse is open source, and self-hosting is a first-class deployment mode: c…
Langfuse is the open-source alternative to Arize AX and Phoenix. September 2026 comparison of unit vs span+GB pricing, evals, and the Dynatrace acquisition.
Language: en
Indexed excerpt
CompareArize / PhoenixCopy pageLangfuse vs. Arize AX / Arize Phoenix This guide compares Langfuse, Arize Phoenix, and Arize AX. All Arize facts were checked against public Arize sources in September 2026. Phoenix and AX are different products. Phoenix is source-available (Elastic License 2.0) on SQLite or PostgreSQL. AX is proprietary on adb (public launch June 2025). Langfuse is one MIT-licensed codebase on ClickHouse (open source since 2016), identical on Cloud and self-hosted. On August 13, 2026, Dynatrace announced a definitive agreement to acquire Arize. The transaction has not closed; this page compares AX and Phoenix as they ship today. See the FAQ. TL;DR Choose Langfuse if you want: 1. The same product in production. Tracing, evaluations, prompt management, and experiments in the MIT distribution. Cloud and self-host share the codebase. Used by 21 of the Fortune 50. 2. An open data plane. Every span lands in ClickHouse, open source since 2016 with 2,000+ contributors. Self-hosters query it in SQL. Cloud and self-host share the same engine. 3. One meter, years of history, code evals on every plan. A unit is a trace, an observation, or a score. Pro includes 3-year retention. …
Query Langfuse data via Python and JS/TS SDKs using the high-performance data APIs.
Language: en
Indexed excerpt
DocsAPI & Data PlatformFeaturesQuery via SDKsCopy pageQuery Data via SDKs Langfuse is open-source and data tracked with Langfuse is open. Use the Python and JS/TS SDKs to query the same public APIs without writing raw HTTP requests. Common use cases: Query row-level observations for evaluation pipelines, few-shot examples, or fine-tuning datasets. Query aggregate cost, usage, latency, volume, and score metrics for dashboards or billing workflows. Programmatically create datasets. If you are new to Langfuse, we recommend familiarizing yourself with the Langfuse data model. New data is typically available for querying within 15-30 seconds of ingestion, though processing times may vary at times. Please visit status.langfuse.com if you encounter any issues. SDKs Via the SDKs for Python and JS/TS you can easily query the API without having to write the HTTP requests yourself. The api namespace is auto-generated from the Public API (OpenAPI). Method names mirror REST resources and support filters and pagination. From Python SDK v4 and JS/TS SDK v5 onward, the high-performance observations and metrics APIs are the defaults: api.observations (formerly api.observations_v_2 / api.observation…
All Langfuse data and features are available via the API. Follow this guide to authenticate, ingest traces, and retrieve observations, scores, and experiments.
Language: en
Indexed excerpt
DocsAPI & Data PlatformFeaturesPublic APICopy pagePublic API Langfuse is open and meant to be extended via custom workflows and integrations. All Langfuse data and features are available via the API. There are 3 different groups of APIs: This page -> Project-level APIs: CRUD traces/evals/prompts/configuration within a project Organization-level APIs: provision projects, users (SCIM), and permissions Instance Management API: administer organizations on self-hosted installations API reference This page covers concepts and workflows. For the complete request and response contract of every endpoint — parameters, schemas, and interactive examples — see the API reference: API Reference: https://api.reference.langfuse.com OpenAPI spec: https://cloud.langfuse.com/generated/api/openapi.yml Quickstart Obtain credentialsThe public and secret keys are available in the Langfuse project settings.Select the regional base URLPathCloud USCloud EUCloud JapanHIPAA US/api/publichttps://us.cloud.langfuse.com/api/publichttps://cloud.langfuse.com/api/publichttps://jp.cloud.langfuse.com/api/publichttps://hipaa.cloud.langfuse.com/api/publicMake an authenticated requestExample:curl -u public-key:secret-key …
Discovered: Last checked: Content changed:
ArticleInternal link
Langfuse v4: up to 165× faster · Read moreLangfuse v4 is here: real-time, up to 165× faster · Read more
Query and evaluate every agent step directly, with initial table loads in milliseconds and at least 10x faster dashboards for large projects.
Language: en
Indexed excerpt
← Back to changelogAugust 17, 2026Langfuse v4 is live: faster at scale, with more ways to search, monitor, and evaluateMax DeichmannQuery and evaluate every agent step directly, with initial table loads in milliseconds and at least 10x faster dashboards for large projects. Langfuse v4 is live on Langfuse Cloud and generally available for self-hosted deployments. It makes it faster to debug, evaluate, and monitor complex LLM applications by letting you work with every LLM call, tool execution, and agent step directly. Initial table loads over large datasets drop from seconds to milliseconds, and dashboards over longer time ranges load at least 10x faster in large projects. Langfuse Cloud becomes v4-only on November 16, 2026 (2026-11-16). Most projects need no migration. If the Migration Assistant shows actions for your project, complete them before this date. New to LangfuseThere is nothing to migrate. Start tracing with the v4 data model.Using Langfuse CloudOpen the Migration status page and complete only the checks detected for your project.Self-hosting LangfusePlan the upgrade on your schedule and review the infrastructure requirements first. What we shipped in Langfuse v4 Langfu…
Retrieve custom metrics from Langfuse for flexible analytics and reporting.
Language: en
Indexed excerpt
DocsMetricsFeaturesMetrics APICopy pageMetrics API GET /api/public/v2/metrics The Metrics API enables you to retrieve customized analytics from your Langfuse data. This endpoint allows you to specify dimensions, metrics, filters, and time granularity to build powerful custom reports and dashboards for your LLM applications. What you can do Use the Metrics API to: Aggregate cost, token usage, volume, latency, and score data. Group results by supported dimensions, such as model or trace attributes. Filter data and analyze trends over time. Power custom reports, dashboards, billing, and monitoring workflows. For supported views, fields, query parameters, response schemas, and interactive examples, see the v2 Metrics API Reference. For practical Python examples, see the Metrics API v2 cookbook. The deprecated GET /api/public/metrics and GET /api/public/metrics/daily endpoints are documented, with migration steps, in Migration of deprecated APIs. Metrics API v2 Where is this feature available?HobbyAvailableCoreAvailableProAvailableEnterpriseAvailableSelf HostedLangfuse v4+ Data availability: Data from older SDKs (langfuse-python < 4.7.0, langfuse-js < 5.4.0) or direct OpenTelemetry expo…
Data model for Datasets, DatasetItems, DatasetRuns, DatasetRunItems, and function definitions for Tasks and Evaluators.
Language: en
Indexed excerpt
DocsEvaluationExperimentsData ModelCopy pageExperiments Data Model This page describes the data model for experiment-related objects in Langfuse. For an overview of how these objects work together, see the Concepts page. For score and score config objects, see the Scores data model. For detailed reference please refer to the Python SDK reference the JS/TS SDK reference the API reference How experiments are created You create experiment runs in Langfuse through one of these paths: PathUse whenExperiments via SDKPython or JS/TS Experiment runnerExperiments via UIPrompt or model experiments from the dataset pageExperiments via OpenTelemetryDirect OTEL ingestion: other languages, custom OTLP pipelines, or re-ingesting experiment traces To read experiment runs, items, and scores after they exist, use the Experiments API. There is no public REST endpoint for creating new experiment runs; the legacy POST /api/public/dataset-run-items path is deprecated. Objects Datasets Datasets are a collection of inputs and, optionally, expected outputs that can be used during Dataset runs. Datasets are a collection of DatasetItems. Dataset object AttributeTypeRequiredDescriptionidstringYesUnique identi…
Group OpenTelemetry traces as Langfuse experiments by attaching experiment and item attributes.
Language: en
Indexed excerpt
DocsEvaluationExperimentsExperiments via OpenTelemetryCopy pageExperiments via OpenTelemetry Attach experiment metadata to OpenTelemetry spans so Langfuse groups those traces as an experiment run. Using Python or TypeScript? Use Experiments via SDK. The Langfuse SDKs set these attributes automatically. See the Experiments data model for how datasets, experiment runs, items, traces, and scores relate. Ingest experiment spans Set experiment and item attributes on your OTEL spans so Langfuse can synthesize them as one experiment. The attribute list, baggage propagation pattern, and per-item loop live in the OpenTelemetry integration docs: Ingest experiment spans with OpenTelemetry After traces land, use the Experiments API to list runs and fetch items.Was this page helpful?GoodBadSupportLast updated on PreviousExperiments in CI/CDNextCompare ExperimentsGitHubDiscordXYouTubeLinkedInProductObservabilityPrompt ManagementEvaluationsMetricsLangfuse for AgentsChat agentsCoding agentsWorkflow automationPlaygroundPricingEnterpriseDevelopersDocumentationSelf-HostingSDKsIntegrationsAPI ReferenceStatusTalk to UsResourcesBlogChangelogEventsRoadmapInteractive DemoCustomersCompareAI Engineering Lib…
Langfuse is designed to be extensible and flexible. People using Langfuse are building all kinds of workflows and customizations on top of it. This is powered by our open data platform.
Language: en
Indexed excerpt
DocsAPI & Data PlatformOverviewCopy pageAPI & Data Platform Langfuse is designed to be open, extensible and flexible (see why Langfuse?). People using Langfuse are building all kinds of workflows and customizations on top of it. This is powered by our open data platform. Example use cases: Billing based on LLM costs tracked in Langfuse Reporting of online evaluations in external dashboards Fine-tuning based on raw exports of traces Correlation of LLM Evals with observed user behavior in Data Warehouse Start here Choose the data access path based on what you want to build: GoalRecommended pathWork with Langfuse from a terminal or coding agentCLIConnect an AI tool that cannot run shell commandsMCP ServerQuery aggregate cost, usage, latency, volume, or score metricsMetrics API v2Retrieve experiment runs, items, and evaluation scoresExperiments APIRetrieve row-level spans, generations, or eventsObservations API v2Use the API from Python or JS/TSQuery via SDKsExport large volumes on a scheduleBlob Storage ExportDownload a filtered one-off exportExport from UIManage prompts, datasets, projects, and other resourcesPublic API Features Langfuse for AgentsCLIMCP ServerPublic APIQuery via SDK…
Langfuse is the open-source LangSmith alternative. September 2026 comparison of MIT self-hosting, ClickHouse vs SmithDB, evals, and unit pricing with no seats.
Language: en
Indexed excerpt
CompareLangSmithCopy pageLangfuse vs. LangSmith This guide outlines the key differences between Langfuse and LangSmith to help engineering teams choose the right AI engineering platform. All LangSmith facts were checked against public LangChain sources in September 2026. TL;DR Choose Langfuse if you want: 1. Control without rebuilding governance. MIT-licensed self-hosting on every tier, including free, with SSO and org RBAC in the OSS distribution, and a framework-agnostic approach built on OpenTelemetry. 2. Data sovereignty. A ClickHouse data plane, open source since 2016, that you can query, join, and hire for. Cloud and self-hosted share the same engine. 3. One usage meter and no seat fees. Traces, observations, and scores bill as units. Langfuse Core and higher plans include unlimited users. Choose LangSmith if you are an all-in LangChain/LangGraph shop that wants a managed SaaS covering observability, evals, and agent deployment infrastructure in one closed platform. Open source and distribution Langfuse is open source (MIT) and self-hosting is a first-class deployment mode. Tracing, evaluations, prompt management, experiments, annotation, and playground ship in the MIT distri…
Langfuse is the open-source alternative to Datadog Agent Observability. September 2026 comparison of self-hosting, unit vs LLM-span pricing, and evals.
Language: en
Indexed excerpt
CompareDatadog Agent ObservabilityCopy pageLangfuse vs. Datadog Agent Observability This guide outlines the key differences between Langfuse and Datadog Agent Observability. All Datadog facts were checked against public Datadog sources in September 2026. TL;DR Choose Langfuse if you want: 1. MIT-licensed self-hosting. Cloud and self-hosted run the same product. Tracing, evaluations, prompt management, and experiments are in the free OSS distribution. Used by 21 of the Fortune 50. 2. Months to years of queryable history. Core is $29/month with 90 days. Pro includes 3 years on ClickHouse. 3. Purpose-built AI quality workflows. Code evaluators on live observations, experiments in the UI and CI, and prompt-version metrics. Choose Datadog Agent Observability if you want APM and LLM tracing in one Datadog deployment. You accept SaaS-only hosting and 15-day default trace retention. Many teams run both: Datadog for APM and infrastructure, Langfuse for AI engineering, from the same OpenTelemetry instrumentation. See Using Langfuse and Datadog together. Open source and distribution Langfuse is open source (MIT). Self-hosting is a first-class deployment mode and uses the same codebase as Lang…
Langfuse now delivers faster product performance at scale. See the overview page for rollout details, access, and migration steps.
Language: en
Indexed excerpt
← Back to changelogMarch 10, 2026Simplify Langfuse for ScaleLangfuse now delivers faster product performance at scale. See the overview page for rollout details, access, and migration steps. Langfuse is rolling out a simplified architecture built for significantly faster product performance at scale. Charts, filters, APIs are becoming much faster across Langfuse Cloud. Open the Langfuse v4 Docs To avoid delays in the new experience and see your data in real time, upgrade to Python SDK v4.7.0+ and JS/TS SDK v5.4.0+. What changes for you The main table in Langfuse is now observations — every LLM call, tool execution, and agent step is a row you can query directly. As agentic applications grow more complex, a single trace can contain thousands of operations, and the interesting ones are rarely at the top level. trace_id works like any other filter column (session_id, user_id, score) to group related observations together. Filter by observation name, type, or model to surface the operations that matter most — then save that view for one-click access. See the guide on working with observations in the next version of Langfuse (v4) for common workflows and saved views to get started. What…