Langfuse

What is Langfuse?

Langfuse is an open-source AI engineering platform for observability and evaluation of LLM applications and agents. It records traces, sessions, users, token usage and cost, then adds prompt management with versioning and a playground, datasets and experiments, LLM-as-judge evaluators, human annotation queues and custom dashboards. Teams send data through Python and JavaScript SDKs, OpenTelemetry, native framework integrations or a LiteLLM proxy, and can reach everything through a public API, an MCP server and a CLI. You can use the managed Langfuse Cloud or self-host under the MIT license with no usage limits. Cloud pricing is based on ingested units, which count traces, observations and scores. The Hobby plan is free with 50k units a month, Core is $29/month, Pro is $199/month, and Enterprise is $2,499/month, with extra units at $8 per 100k.

Use cases

• Trace LLM application and agent runs
• Track token usage and cost per user
• Version and test prompts in a playground
• Run LLM-as-judge evaluations on outputs
• Collect human annotations on traces
• Self-host an LLM observability stack

Who it's for

• AI engineers
• ML platform teams
• Product teams
• Enterprise AI teams

Why use Langfuse

• Tracing, prompt management and evaluation live in one open platform, so teams do not stitch together separate tools.
• The free Hobby plan needs no credit card, and the MIT-licensed self-hosted version has no usage limits.
• Multiple ingestion paths, including OpenTelemetry and a LiteLLM proxy, fit many stacks.
• Higher plans extend data access to 3 years and add SOC2 and ISO27001 reports.
• Discounts are available for startups, research and students, and open-source projects.

Langfuse features

• Traces and graphs for agents, with session and user tracking
• Token and cost tracking
• Python and JavaScript SDKs, OpenTelemetry and LiteLLM proxy logging
• Prompt versioning, release management, composability and caching
• Playground and prompt experiments
• Datasets, experiments via SDK or UI, and custom evaluation scores
• LLM-as-judge evaluators and human annotation queues
• Custom dashboards and alerts
• Public API, MCP server and CLI
• Batch export plus PostHog and Mixpanel integrations

Langfuse integrations

• LiteLLM
• OpenTelemetry
• PostHog
• Mixpanel
• Slack
• ClickHouse
• Okta
• AWS
• Azure
• GCP

How to get started with Langfuse

1. Sign up for Langfuse Cloud or deploy the open-source version with Docker Compose or Helm.
2. Instrument your app with an SDK, OpenTelemetry or a native framework integration so traces flow in.
3. Manage prompts, build datasets and attach evaluators or human annotations to score outputs.
4. Use dashboards, alerts and the API to monitor quality, cost and latency over time.

Langfuse limitations

• Hobby plan includes 50k units a month, 30 days of data access and 2 users
• Cloud usage is billed on ingested units, and units created by features like LLM-as-a-Judge also count
• Teams add-on features such as SSO enforcement need the paid Teams add-on on Pro
• Some alert and API features in self-hosted setups require Langfuse v4 or later

Langfuse pricing

Yes, Langfuse has a free plan. Paid plans start at $29 per month, according to the official pricing page on 24 Oct 2026.

Langfuse plans

• Hobby

  Free

  • 50k units / month included
  • 30 days data access
  • 2 users
  • Community support via GitHub

• Core

  $29/month

  • 100k units / month included
  • Additional units $8/100k
  • 90 days data access
  • Unlimited users
  • In-app support

• Pro

  $199/month

  • 100k units / month included
  • 3 years data access
  • Data retention management
  • Unlimited annotation queues
  • SOC2 and ISO27001 reports

• Enterprise

  $2,499/month

  • Audit logs
  • SCIM API
  • Custom rate limits
  • Uptime SLA
  • Named lead support engineer

• Teams Add-on

  $300/month

  optional add-on to Pro

  • Enterprise SSO
  • SSO enforcement
  • Fine-grained RBAC
  • Dedicated Slack channel

• Self-Hosted Open Source

  Free

  • MIT License
  • Unlimited units
  • Helm chart and deployment docs

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

Langfuse alternatives

The main Langfuse alternatives listed on Debuly are Arize Phoenix, Helicone, Braintrust, Arize AX, VoltAgent and Opik. The table compares pricing and platforms; order is A–Z.

1. 

   Arize AX

   Observability and evaluation platform for AI agents and LLM apps, with tracing, issue detection and experiments on a free plan.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Arize AX (official site)
2. 

   Arize Phoenix

   Open-source tool for tracing, evaluating and experimenting on LLM apps and agents, free to self-host or run locally.

   Developer platformsAutomation and agents
   Open sourceVisit Arize Phoenix (official site)
3. 

   Braintrust

   AI observability and evaluation platform for tracing LLM apps and agents, scoring outputs and catching regressions before release.

   Developer platformsCoding
   FreemiumVisit Braintrust (official site)
4. 

   Helicone

   Open-source LLM observability and AI gateway that logs requests, tracks cost and latency, and routes to 100+ models through one API.

   Developer platformsCoding
   FreemiumVisit Helicone (official site)
5. 

   Opik

   Comet's open-source platform for tracing, evaluating and monitoring LLM applications and agents, self-hosted or in the cloud.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Opik (official site)
6. 

   VoltAgent

   Open-source TypeScript framework for building AI agents, paired with the VoltOps console for tracing, evals and deployment.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit VoltAgent (official site)

Langfuse FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from langfuse.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.