Langfuse
Trace, evaluate and debug LLM agents.
In short
- What it is: Observability, 12k+ GitHub stars, licensed MIT core.
- Replaces: LLM observability SaaS — around $1,188/year saved.
- Installed for: $199 on your own server — you own it, no monthly platform fee.
- Licence: MIT core. MIT core resellable.
$1,188
replaces per year
$5,940
saved over 5 years
$199
one-time install (yours forever)
48h
delivery
Langfuse vs the tools it replaces.
| Feature | Commercial tool | Langfuse (self-hosted) |
|---|---|---|
| Pricing | Usage | Flat |
| Data | Vendor | Yours |
| Licence | Proprietary | MIT |
Replaces
- LLM observability SaaS — ~$99/mo
Licence posture
MIT core. You can white-label and resell this. We'll brand it as yours.
MIT core resellable.
Turn it into a business.
Sell "we monitor and improve your AI" as part of a care plan. Self-host the traces so client data stays put.
The install, step by step.
Docker Compose
Postgres
Keys
Dashboards
Alerts
Watch a Langfuse install.
We generate an architecture diagram of the repo and walk through the deployment. Enter your name and email to unlock the demo.
Install Langfuse — $199.
Flat fee, delivered in 24–48 hours, on your own server, with HTTPS, backups and the AI keys wired.
Langfuse: common questions
What does Langfuse replace?
LLM observability SaaS — about $1,188 a year in subscriptions you stop paying.
How much does it cost to run Langfuse?
No licence fee — it ships under MIT core. You pay only for your own server and any model/API keys. We install, harden and hand it over for a one-time $199.
Can I white-label and resell Langfuse?
You can white-label and resell this. We'll brand it as yours. MIT core resellable.
How hard is the install, and how long does it take?
We rate it medium. Delivered in 24–48 hours on your own server, with HTTPS, backups and the AI keys wired.
What do I actually own after installing Langfuse?
The running app, your data and the brand. No monthly platform fee and no per-seat pricing — you own it outright.
Other repos people install with Langfuse.
Phoenix
Arize's open-source LLM tracing and evals.
Apache-2.0 · 5k+★
ObservabilityHelicone
An LLM observability proxy with caching.
Apache-2.0 · 3k+★
ObservabilityPromptfoo
Test and red-team your prompts and models.
MIT · 6k+★
ObservabilityDeepEval
Unit testing for LLM outputs.
Apache-2.0 · 5k+★
ObservabilityEvidently
Monitor and evaluate ML/LLM systems.
Apache-2.0 · 6k+★
Chat UILibreChat
A full multi-model ChatGPT-style interface with auth, multi-user and presets.
MIT · 40k+★