Stack

Langfuse moved off homegrown billing onto Stripe Billing while keeping an in-house ClickHouse meter

Langfuse pricing

Langfuse's Stripe case study discloses that it started with an internal billing system based on a simple per-event flat rate, then migrated to Stripe Billing (usage-based), Stripe Checkout, and Stripe Tax for the rails. It keeps the metering in-house: it posts hourly event counts from its own ClickHouse/OLAP data store to Stripe's metered-usage API, handling up to 200M events/month per account — the operational mirror of its "unit = trace + observation + score" price metric. A "Senior Backend Engineer (IAM and Billing)" req (open in the 2026-04-17 ATS snapshot, since closed) staffed that billing platform; growth analytics run on a bought PostHog/BigQuery/dbt/Metabase stack.

About Langfuse
langfuse.com ↗

Langfuse is an open-source LLM engineering platform (observability, evals, prompt management) acquired by ClickHouse in January 2026; it can be self-hosted free under MIT license or run as Langfuse Cloud.

Pricing model freemiumhybridsubscription
Billing units unitseventsseats
Sales motion self servesales led
Free tier
Yes
Commits
Available
Transparency
public
Full Langfuse breakdown Plans, meters, and history All pricing activity Track changes across AI companies

Langfuse pricing history

  1. Jan 2026
    Acquired by ClickHouse at $15B; no pricing change
  2. Jan 2025
    Unit-based metering: traces + observations + scores
  3. Sep 2024
    v3 migrates core data layer to ClickHouse
  4. Aug 2023
    Launched as open-source LLM tracing (YC W23)
Full Langfuse timeline
All pricing activity