Langfuse Blog
The latest updates from Langfuse. See Changelog for more product updates.
Highlights
engineering
How we built Langfuse Halo
How we built Langfuse Halo, our powerful in-app assistant, on our existing MCP server, why the agent runs outside its sandbox, and how we handle background runs.
Oct 21, 2026
ยทBen
engineering
Catching conversation signals in Langfuse
Score every chat agent conversation for frustration, corrections, follow-ups, and resolution โ in one Jev pass, at a fraction of LLM-as-a-judge cost.
Sep 23, 2026
ยทAnnabell
engineering
Using TypeSafe's Jev for evals
A specialized decision model for routing, classification, and eval verdicts โ and how to score production traces with Jev-as-a-judge in Langfuse.
Sep 18, 2026
ยทAnnabell, Hassieb
All Posts
engineering
We launched Halo ๐
We launched Halo, the assistant inside Langfuse. It can investigate traces, help set up evals, build dashboards, and take action with your approval.
Oct 23, 2026
ยทLotte
guide, architecture
Operating Langfuse as an internal service
Coordinate SDK upgrades, monitor service health, and support evaluation workflows, access reviews, and cost allocation across application teams.
Oct 22, 2026
ยทBoris, Doneyli
guide, architecture
Provisioning and governing Langfuse across teams
Choose project boundaries, provision access, and verify safe, useful telemetry as more application teams adopt Langfuse.
Oct 22, 2026
ยทBoris, Doneyli
update
Langfuse September Update
Jev-as-a-judge, Halo sandbox, rebuilt experiment results, evaluator backfills and more
Sep 30, 2026
ยทMarc
update
Langfuse August Update
Langfuse v4, new evaluator setup flow, multi-modal evaluators, CLI 1.0, responsive timeline and more
Aug 31, 2026
ยทMarc
update
Langfuse July Update
Chart any table, dashboards via API, CLI, and MCP, new graph view modes, media previews and more
Jul 31, 2026
ยทMarc
engineering
Yes, you can copy our eval setup
Evals need to be custom to your app. At least some of them. A good chunk you can, and should be copying. Here is our setup for our docs chatbot.
Jul 22, 2026
ยทAnnabell
guide
Building Deployment Gates for LLMs and AI Agents in Financial Services
How we used Langfuse datasets, experiments, and the public API to build automated pass/fail deployment gates for LLMs and AI agents in financial services.
Jul 15, 2026
ยทDoneyli
engineering
Designing the runtime for Langfuse code evaluators
Code evaluators let you score traces with your own Python or TypeScript code. A look at the execution model behind them: the requirements, the options we rejected, and the security stance we adopted.
Jun 22, 2026
ยทTobias
engineering
AI is eating the AI engineering loop
The full AI engineering loop can technically be automated now. But that doesn't mean it should. Here is what we think you should hand to agents, and what you should keep doing yourself.
Jun 9, 2026
ยทLotte
engineering
How we use agents to review production infrastructure
How repo-owned agent workflows help us review incidents, infra cost, security findings, and bugs in production.
Jun 5, 2026
ยทMax
update
Langfuse May Update
Code Evaluators, full-text search, Langfuse MCP, Experiments in CI/CD and more
May 31, 2026
ยทMarc
update
Langfuse April Update
Japan Cloud Region, new Experiments, Langfuse Academy, LLM-as-a-Judge API and more
Apr 30, 2026
ยทMarc
announcement
Langfuse Cloud ๆฅๆฌใชใผใธใงใณใ้ๅงใใพใใ
Langfuse Cloud ๆฅๆฌใชใผใธใงใณใๅ ฌ้ใใพใใใLLM ใฎใใฌใผในใ่ฉไพกใใผใฟใๆฅๆฌๅฝๅ ใซไฟ็ฎกใใใใใผใ ๅใใฎๅฐ็จใฏใฉใฆใใชใผใธใงใณใงใใ
Apr 27, 2026
ยทMarc, Clemens, Max
announcement
Langfuse Cloud Japan Region
Langfuse Cloud Japan is live. A dedicated cloud region hosted in Japan for teams that need their LLM observability data to stay in Japan.
Apr 27, 2026
ยทMarc, Clemens, Max