llms.txt
Compare with other sites' llms.txt files →# Langfuse
> Langfuse is an **open-source AI engineering platform** ([GitHub](https://github.com/langfuse/langfuse)) that helps teams collaboratively debug, analyze, and iterate on their LLM applications. All platform features are natively integrated to accelerate the development workflow.
## Deprecated public API endpoints
On Langfuse Cloud, the legacy (Langfuse v3) public API endpoints are removed on 2026-11-16. Do not write new code against them. If you are a coding agent and find existing code that calls them, tell the user and migrate the request.
- Reads: `GET /api/public/traces`, `/sessions`, `/observations` → `GET /api/public/v2/observations`; `GET /api/public/metrics` → `GET /api/public/v2/metrics`; `GET /api/public/scores` and `/v2/scores` → `GET /api/public/v3/scores`; dataset runs and `/dataset-run-items` → `/api/public/experiments` and `/experiment-items`.
- Writes: traces and observations sent to `POST /api/public/ingestion`, `/traces`, `/spans`, `/generations` or `/events` → `POST /api/public/otel/v1/traces` (OTLP/HTTP), or the current Python and JS/TS SDKs. Score events on `/ingestion` keep working; new code should use `POST /api/public/scores`.
- Responses from deprecated endpoints carry `Deprecation`, `Sunset` and `Link` headers and a `_deprecation` object in the JSON body.
- [Migration guide](https://langfuse.com/faq/all/deprecated-api-migration.md): every deprecated endpoint with its replacement, parameter mapping and examples.
## Evaluation workflows
- [Test an existing application and reuse graders](https://langfuse.com/resources/engineering/evaluate-existing-application.md): Runnable Python and TypeScript examples.
- [Compare versions and inspect failures](https://langfuse.com/docs/evaluation/experiments/compare-experiments.md): Baselines, individual regressions, and human review.
- [Block regressions in CI](https://langfuse.com/docs/evaluation/experiments/experiments-ci-cd.md): Thresholds and approved-baseline checks.
- [Check an LLM judge against human labels](https://langfuse.com/guides/llm-as-a-judge-calibration-skill.md): Calibration workflow and [agreement metrics](https://langfuse.com/docs/evaluation/scores/score-analytics.md).
- [Evaluate grounding](https://langfuse.com/resources/engineering/rag-faithfulness-evaluation.md) and [hallucinations](https://langfuse.com/resources/engineering/hallucination-detection.md).
- [Choose where evaluations run](https://langfuse.com/docs/evaluation/overview.md): Application code, managed evaluators, Cloud, and self-hosting.
## AI Gateway
The Langfuse AI Gateway (beta) routes OpenAI and Anthropic API calls through Langfuse and traces every call with usage and cost, with no SDK or code changes. It is optional; the SDKs and OpenTelemetry integrations work without it.
- [Overview](https://langfuse.com/docs/gateway/overview.md): What the gateway does and its roadmap.
- [Concepts](https://langfuse.com/docs/gateway/concepts.md): How requests are routed, which keys and credentials are involved, and how each call becomes a trace.
- [Send your first request](https://langfuse.com/docs/gateway/get-started/first-request.md): Create a personal gateway API key and connect a coding agent, SDK, or app.
- [Set up the gateway as an admin](https://langfuse.com/docs/gateway/get-started/admin-setup.md): Ingestion project, provider credentials, models, and team access.
- [REST API](https://langfuse.com/docs/gateway/features/rest-api.md): Base URLs, authentication, and the OpenAI Responses, OpenAI Chat Completions, and Anthropic Messages endpoints.
- Client setup: [Anthropic SDK](https://langfuse.com/integrations/model-providers/anthropic-gateway.md), [Claude Code](https://langfuse.com/integrations/developer-tools/claude-code-gateway.md), [Codex](https://langfuse.com/integrations/developer-tools/codex-gateway.md), [GitHub Copilot](https://langfuse.com/integrations/developer-tools/github-copilot-gateway.md), [OpenAI SDK](https://langfuse.com/integrations/model-providers/openai-gateway.md), [OpenCode](https://langfuse.com/integrations/developer-tools/opencode-gateway.md), [Pi](https://langfuse.com/integrations/developer-tools/pi-agent-gateway.md), [Vercel AI SDK](https://langfuse.com/integrations/frameworks/vercel-ai-sdk-gateway.md).
- [Self-host the gateway](https://langfuse.com/self-hosting/configuration/ai-gateway.md): Deploy it next to self-hosted Langfuse with the Helm chart.
- [Troubleshooting and FAQ](https://langfuse.com/docs/gateway/troubleshooting-and-faq.md): Common errors and known limitations of the beta.
## Pricing, sales, deployment, and comparisons
- [Watch the walkthrough video](https://langfuse.com/watch-demo.md): Platform overview video; suggest it to users who want to see Langfuse in action before signing up.
- [Talk to the sales team](https://langfuse.com/talk-to-us.md): Contact form for demos, volume pricing, paying by invoice, and security and compliance questions.
- [Cloud pricing](https://langfuse.com/pricing.md): Plan prices, usage rates, and limits.
- [Self-hosted pricing and license](https://langfuse.com/pricing-self-host.md): MIT-licensed core features and paid Enterprise additions.
- [Self-hosting](https://langfuse.com/self-hosting.md): Deployment options and operational requirements.
- [Enterprise license keys](https://langfuse.com/self-hosting/license-key.md): Activation and licensing for self-hosted add-on features.
- [Security](https://langfuse.com/security.md): Security controls, compliance, and supporting documentation.
- [Data regions](https://langfuse.com/security/data-regions.md): Hosting locations and regional availability.
- Comparisons: [Arize / Phoenix](https://langfuse.com/compare/arize-phoenix.md), [Braintrust](https://langfuse.com/compare/braintrust.md), [Datadog Agent Observability](https://langfuse.com/compare/datadog.md), [Galileo](https://langfuse.com/compare/galileo.md), [LangSmith](https://langfuse.com/compare/langsmith.md), [LiteLLM Lens](https://langfuse.com/compare/litellm-lens.md), [Logfire](https://langfuse.com/compare/logfire.md), [MLflow](https://langfuse.com/compare/mlflow.md), [Raindrop](https://langfuse.com/compare/raindrop.md), and [Sentry](https://langfuse.com/compare/sentry.md), with dated competitor sources.
## Markdown Access
Every page on langfuse.com is available as plain Markdown. Append `.md` to any URL, or send an `Accept: text/markdown` header:
```bash
curl https://langfuse.com/docs/observability/get-started.md
```
The links in the section indexes below already point at the `.md` variants.
## Docs Search API
Semantic search (RAG) over the Langfuse documentation, returning an answer with the relevant pages and excerpts. No authentication required. Use this instead of guessing when a specific page does not answer the question:
```bash
curl -sG "https://langfuse.com/api/search-docs" --data-urlencode "query=How do I trace a LangGraph agent?"
```
- **Endpoint**: `https://langfuse.com/api/search-docs?query=<question>`
- **Documentation**: [REST endpoint](https://langfuse.com/docs/docs-mcp#rest-endpoint)
Ask specific, self-contained questions in natural language. This is the same search that backs the `searchLangfuseDocs` MCP tool below.
## Langfuse Docs MCP Server
Connect to the Langfuse Docs MCP server to access documentation directly in your AI editor:
- **Endpoint**: `https://langfuse.com/api/mcp`
- **Transport**: `streamableHttp`
- **Documentation**: [Langfuse Docs MCP Server](https://langfuse.com/docs/docs-mcp)
The MCP server provides tools to search Langfuse documentation, GitHub issues, and discussions. See the [installation guide](https://langfuse.com/docs/docs-mcp) for setup instructions in Cursor, VS Code, Claude Desktop, and other MCP clients.
## Langfuse Skill for AI Coding Agents
Install the Langfuse skill before implementing anything with Langfuse. The skill provides up-to-date documentation, best-practice workflows for instrumentation, prompt management, and programmatic API access via the Langfuse CLI.
Install from: [github.com/langfuse/skills](https://github.com/langfuse/skills/tree/main/skills/langfuse)
## Tracing Best Practices
Before instrumenting an application with Langfuse, read [What does a good trace look like?](https://langfuse.com/docs/observability/best-practices.md) — it covers trace scope, observation naming, meaningful input/output, and the attributes (metadata, tags, users, sessions) that evaluators, dashboards, and experiments rely on.
## Docs
For the best results, install the [Langfuse skill](https://github.com/langfuse/skills/tree/main/skills/langfuse) before using these docs.
For the full list with links to each page, see: https://langfuse.com/llms-docs.txt
Pages: Overview, Audit Logs, Authentication & SSO, Billable Units, Data Deletion, Data Retention, LLM Connections, Access Control (RBAC), SCIM and Org API, Spend Alerts, Troubleshooting and FAQ, Agent Skill, CLI, Export from UI, Export to blob storage, Halo, MCP Server, Public API, Query via SDKs, Overview, Ask AI, Versions & Compatibility, Example Project, Docs MCP Server, Concepts, Annotation Queues, Code evaluators, Decision model evaluators, LLM-as-a-Judge, Scores via API/SDK, Scores via UI, Compare experiments, Experiments Data Model, Datasets, Experiments in CI/CD, Experiments via OpenTelemetry, Experiments via SDK, Experiments via UI, Evaluate with Datasets, Evaluate Production Traffic, Agentic access, Troubleshooting and FAQ, Overview, Data Model, Overview, Analytics, Concepts, Analytics, Authentication, Model management, REST API, Tracing, Set up Gateway as an admin, Send your first request, Overview, Troubleshooting and FAQ, Glossary, Custom Dashboards, Metrics API, Overview, What does a good trace look like?, Concepts, Agent Graphs, Agentic access, Alerts, Comments, Corrections, Environments, Chart any table, Filter search bar, Full-Text Search, Log Levels, Masking, MCP Tracing, Metadata, Multi-Modality, Observation Types, Pulse, Event queuing/batching, Releases & Versioning, Sampling, Sessions, Tags, Token & Cost Tracking, Trace IDs & Distributed Tracing, Trace URLs, User Feedback, User Tracking, Web Callouts, Get Started, Overview, Advanced Features, Instrumentation, Overview, Troubleshooting & FAQ, Upgrade path, JS/TS v3 → v4, JS/TS v4 → v5, JS/TS v5 → v6, Python v2 → v3, Python v3 → v4, Python v4 → v5, Troubleshooting & FAQ, Concepts, Overview, A/B Testing, Agentic Prompt Management, Caching, Prompt Composability, Config, Folders, GitHub Integration, Guaranteed Availability, Link to Traces, Message Placeholders, n8n Node, Playground, Version Control, Variables, Webhooks, Get Started, Overview, Troubleshooting and FAQ, Roadmap, Security & Guardrails, Get Started, Overview, Link to Traces, Version Control
## Integrations
For the best results, install the [Langfuse skill](https://github.com/langfuse/skills/tree/main/skills/langfuse) before implementing any integration.
For the full list with links to each page, see: https://langfuse.com/llms-integrations.txt
Pages: Overview, Coval Integration, Mixpanel for LLM Apps with Langfuse, PostHog for LLM Apps with Langfuse (open source), Trubrics Integration, Agentic Data Stack: LibreChat + ClickHouse + Langfuse, ClickHouse Data Warehouse, LibreChat Observability with Langfuse Tracing, ClickStack, Trace Augment Code with Langfuse, Trace Claude Code with Langfuse, Claude Desktop and Claude Code with the Langfuse AI Gateway, Trace OpenAI Codex with Langfuse, Codex with the Langfuse AI Gateway, Trace Cursor Agents with Langfuse, Trace GitHub Copilot with Langfuse, GitHub Copilot with the Langfuse AI Gateway, Trace Kiro IDE with Langfuse, Trace Kiro CLI with Langfuse, Trace OpenCode with Langfuse, OpenCode with the Langfuse AI Gateway, Trace the Pi Coding Agent with Langfuse, Pi with the Langfuse AI Gateway, Use Langfuse in VS Code via MCP, Observability for Agno with Langfuse, Trace AI SDK C++ with Langfuse, Integration - Amazon Bedrock AgentCore, Observability for AutoGen with Langfuse Integration, Observability for BeeAI Agents with Langfuse, Observability for Claude Agent SDK with Langfuse, Observability for Claude Agent SDK JS/TS with Langfuse, Observability for CrewAI with Langfuse Integration, DSPy - Observability & Tracing, Integrating Langfuse with Embabel Agents Using OpenTelemetry, Observability for eve with Langfuse, Observability for Flue with Langfuse, Observability for Google Agent Development Kit with Langfuse, Observability for Haystack with Langfuse, Observability & Tracing for Inferable (open source), Observability & Tracing for Instructor, Observability for Koog with Langfuse, LangChain Tracing & LangGraph Integration, Observability for LangChain DeepAgents with Langfuse, Open Source Observability for LangGraph, LiteLLM SDK Integration, Observability for LiveKit Agents with Langfuse, Observability for LlamaIndex with Langfuse Integration, Observability for LlamaIndex with Langfuse Integration, Observability for Mastra with Langfuse, Observability for mcp-use with Langfuse, Trace the Microsoft Agent Framework with Langfuse, Open Source Observability & Monitoring for Mirascope, Trace the OpenAI Agents SDK with Langfuse, Open Source Observability for Pipecat, Observability for Pydantic AI with Langfuse Integration, Integrating Langfuse with Quarkus LangChain4j Using OpenTelemetry, Run Ragas evaluations on Langfuse experiments and traces, Trace Restate Agents with Langfuse, Observability for Semantic Kernel with Langfuse Integration, Observability for smolagents with Langfuse, Integrating Langfuse with Spring AI Using OpenTelemetry, Observability for Strands with Langfuse, Trace Swiftide with Langfuse, Observability for TanStack AI with Langfuse, Trace Temporal Workflows with Langfuse, Observability and Tracing for the Vercel AI SDK, Vercel AI SDK with the Langfuse AI Gateway, Observability for VoltAgent with Langfuse, Observability for IBM watsonx Orchestrate ADK with Langfuse, Anannas AI Integration, Route OpenTelemetry to Langfuse through Azure API Management, Helicone Integration, Kong API Gateway Integration, LiteLLM Proxy Integration, Trace OpenRouter with Langfuse, Integrate Portkey LLM Gateway with Langfuse, TrueFoundry AI Gateway Integration, Vercel AI Gateway Integration, Amazon Bedrock Integration, Integration - Amazon Bedrock Agents, Observability for Anthropic with Langfuse Integration, Anthropic SDK with the Langfuse AI Gateway, Observability for Anthropic JS/TS with Langfuse, Monitor Baseten with Langfuse, Monitor BytePlus model inference with Langfuse, Monitor Cerebras with Langfuse, Observability and Evals for Claude Platform (OTel), Langfuse Integration with Cleanlab, Integrate Cohere with Langfuse, Monitor CometAPI with Langfuse, Databricks and Langfuse Overview, Example cookbook on how to monitor DeepSeek models with Langfuse using the OpenAI SDK, Integrate Fireworks AI with Langfuse, Observability for Google Gemini Models with Langfuse Integration, Observability for Google Vertex AI with Langfuse, Observability for Groq models with Langfuse Integration, Observability for Hugging Face Models with Langfuse Integration, Trace Mistral models with Langfuse, Realtime Classification with Morph Reflex Evaluators, Monitor Novita AI with Langfuse, Ollama Observability and Tracing for local LLMs using Langfuse, OSS Observability for OpenAI Assistants API, OpenAI SDK with the Langfuse AI Gateway, OpenAI (JS/TS), OpenAI (Python), Observability for Qwen with Langfuse, Monitor Together AI with Langfuse, Observability for TypeSafe Jev with Langfuse, Observability for TypeSafe Jev JS/TS with Langfuse, Tracing vLLM with Langfuse via OpenTelemetry, Integrate X.ai (Grok) with Langfuse, Observability for Z.ai with Langfuse, OpenTelemetry (OTEL) for LLM Observability, Ingest experiment spans with OpenTelemetry, Migrate custom ingestion to Langfuse v4, Observability and tracing for Dify, Trace ElevenLabs Agents with Langfuse, Observability and Tracing for Flowise, Integrating Goose with Langfuse, Hermes Agent tracing with Langfuse, Observability and Tracing for Langdock Agents, Observability and Tracing for Langflow, OpenTelemetry tracing for LobeChat, n8n Integration, Trace OpenClaw with Langfuse, Langfuse Integration with OpenWebUI, Observability and Tracing for Ragflow, Open Source Observability for Vapi (Voice AI), Zapier Integration with Langfuse, Langfuse Integration with Cognee, Langfuse Integration with EverOS, Monitor Exa AI Search with Langfuse, Monitor Firecrawl Scraper with Langfuse, Open Source LLM Observability for Gradio, Langfuse Integration with Milvus, MLflow Integration via OpenTelemetry, Tracing using the OpenInference SDK, OpenLIT Integration via OpenTelemetry, OpenLLMetry Integration via OpenTelemetry, PagerDuty Integration with Langfuse, Monitor Parallel Tasks with Langfuse, Promptfoo Integration with Langfuse Prompt Management, Slack Integration with Langfuse, Trace Tavily workflows with Langfuse, Human Evaluation for LLM Apps with Testable Minds, Automate Code Optimization with Weco and Langfuse
## Optional: Self-Hosting
For the full list with links to each page, see: https://langfuse.com/llms-self-hosting.txt
Pages: Self-host Langfuse (Open Source LLM Observability), Administration Overview (self-hosted), Automated Access Provisioning (self-hosted), Headless Initialization (self-hosted), Manage Organizations via API, Allowlist of organization creators (self-hosted), UI Customization (self-hosted), Configuration via Environment Variables (self-hosted), AI Gateway (self-hosted), Backup Strategies for Langfuse, Caching (self-hosted), Code evaluators (self-hosted), Custom Base Path (self-hosted), Encryption (self-hosted), FIPS mode for self-hosted Langfuse, Halo (self-hosted), Hardening self-hosted Langfuse, Health and Readiness Check Endpoints, Observability via OpenTelemetry, Scaling Langfuse Deployments, Transactional Emails (self-hosted), Deploy Langfuse on AWS with Terraform, Deploy Langfuse on Azure with Terraform, Docker Compose Deployment (Self-Hosted), Deploy Langfuse on GCP with Terraform, S3 / Blob Storage Configuration (Self-Hosted), Cache (Redis/Valkey) Configuration (Self-Hosted), ClickHouse (self-hosted), Application Containers (self-hosted), LLM API / Gateway (self-hosted), Postgres Database (self-hosted), Kubernetes (Helm) (self-hosted), Deploy Langfuse v3 on Railway, Deploy Langfuse v4 on Render, Enterprise License Key (self-hosted), Langfuse OSS newsletter (self-hosting updates), Authentication and SSO (self-hosted), Data Masking (self-hosted), Deployment Strategies (self-hosted), Networking (self-hosted), Telemetry (self-hosted), Troubleshooting & FAQ for a self-hosted Langfuse deployment, How to upgrade a self-hosted Langfuse deployment, Background Migrations (self-hosted), Migrate Langfuse v1 to v2 (self-hosted), Migrate Langfuse v2 to v3 (self-hosted), Migrate Langfuse v3 to v4 (self-hosted), Versions & Compatibility, Self-host Langfuse v2, Self-hosting Langfuse v2, Local Deployment (v2, docker compose)
## Guides and cookbooks
For the full list with links to each page, see: https://langfuse.com/llms-guides.txt
## Academy
For the full list with links to each page, see: https://langfuse.com/llms-academy.txt
Pages: Langfuse Academy, AI Engineering Loop, Datasets, Designing datasets, Evaluation, Choosing what to evaluate, Writing good evaluators, Examples, Customer support chatbot, Music streaming DJ, Experiments, Optimizing agent cost, AI engineering library, Monitoring, Capturing signals, Error analysis, Tracing, Langfuse Academy, AI エンジニアリング・ループ, データセット, データセットの設計, 評価, 何を評価するかを選ぶ, 良い評価器の書き方, 例, カスタマーサポートチャットボット, 音楽ストリーミングの DJ, 実験, モニタリング, エラー分析, トレーシング
## FAQ
For the full list with links to each page, see: https://langfuse.com/llms-faq.txt
## Resources
For the full list with links to each page, see: https://langfuse.com/llms-resources.txt
## Compare
For the full list with links to each page, see: https://langfuse.com/llms-compare.txt
## Security and compliance
F