Fully Managed Langfuse
as a Service

Deploy Langfuse as a fully managed service starting at €9/mo. Get automated backups, SSL, updates, support and monitoring included.

Langfuse is an open-source platform for building with language models. It records what every model call, tool step and retrieval actually did, keeps prompts under version control, and scores outputs with evaluators. It gives you the observability layer hosted products sell per seat, on infrastructure you control.

✓Free 7-day trial  ✓99.9% Uptime SLA  ✓No credit card  ✓Cancel anytime

✓Free 7-day trial  ✓99.9% Uptime SLA
✓No credit card  ✓Cancel anytime

Langfuse

Langfuse

STARTING AT

€9/month
Automated Backups
Monitoring
Automated Updates
Auto SSL

USAGE

Unlimited
Human Support
Custom Domains
Terminal Access
File Manager Access
Deploy in your region — 21 locations worldwide
GermanyFinlandNetherlandsUKSwedenUnited StatesCanadaSingaporeJapanAustraliaBrazilSouth Africa+9 more →
Langfuse Preview Image

ABOUT THE SOFTWARE

What is Langfuse

Langfuse is an open-source platform for language-model engineering. It traces model calls, tool steps and retrievals, versions your prompts, scores outputs with evaluators, and turns the result into dashboards. Teams run it as the observability layer under agents and retrieval pipelines.

The project is built by Langfuse GmbH in Berlin with a second entity in San Francisco, came out of Y Combinator, and was acquired by ClickHouse, which reaffirmed the open licence and the self-hosting path. More than three hundred people have contributed code to it, and the SDKs are pulled millions of times a month.

The licensing is worth understanding before you commit to it. The core is MIT: tracing, prompt management, evaluations including model-graded scoring, the playground, annotation queues, datasets, dashboards and single sign-on all sit under the permissive licence. A smaller set of platform features lives in separate enterprise directories that need a paid key to run. We operate the MIT core and leave that key unset, so your instance carries no second licence obligation to anyone.

FEATURES

What Langfuse does

Langfuse covers the loop between shipping a prompt and knowing whether it worked. Tracing shows what happened, evaluations say whether it was any good, prompt management controls what goes out next, and datasets let you check a change before it reaches anyone.

Tracing

Every model call, tool invocation and retrieval step lands in one hierarchical trace. Filter by user, session, latency, cost or your own metadata.

Prompt management

Prompts are versioned and labelled, and the SDK caches them client-side with background revalidation, so a prompt change does not need an application deploy.

Evaluations

Score outputs with model-graded evaluators, deterministic code checks, captured user feedback or manual labelling, attached to the traces they came from.

Datasets and experiments

Build test sets from real traces, then run a candidate prompt or model against them and compare before the change reaches production traffic.

Playground

Open a bad trace in the playground with its inputs intact, adjust the prompt or the model, and rerun it without rebuilding the request by hand.

Annotation queues

Route traces to a human reviewer against a defined rubric, so expert judgement is captured as structured scores instead of a spreadsheet.

Dashboards and metrics

Cost, latency, volume and score trends over any filter, in the interface and through a metrics API for pulling into your own reporting.

OpenTelemetry ingestion

The OTLP endpoint accepts spans from instrumentation you already run, so services outside the Langfuse SDK land in the same trace.

WHAT'S ALWAYS INCLUDED

Every app. Fully managed.
Nothing extra to pay for.

Every app on DANIAN ships with the same operational baseline. You do not choose a backup add-on or a monitoring tier. Hosting, patching, backups, SSL, and support are part of the €9, and they apply to Langfuse the same way they apply to everything else in the catalog.

Automatic updates and patches

Apps run the latest stable version. Security patches applied silently, with rollback if needed.

Daily off-site backups

Multiple daily backups in redundant off-site locations. One-click restore if anything goes wrong.

24/7 uptime monitoring

Continuous monitoring with instant alerting. We respond before you notice.

SSL, firewall, DDoS protection

Auto-renewing SSL, hardened firewall rules, DDoS mitigation on every deployment.

Performance and scaling

We monitor resource usage continuously. When your app needs more headroom, we flag it and upgrade with your explicit approval.

Dedicated engineering support

Real engineers by email and ticket. DNS, SMTP & migration help. All included in €9.

WHY MANAGED

Why teams pick managed Langfuse

The hosted alternative most teams start on raised its usage-based trace pricing and shortened its longer retention window, while keeping a per-seat charge on top. Self-hosting removes both charges and hands you a five-service stack to operate. Managed Langfuse is the third option.

Running Langfuse is not running a container. A production install is Postgres for accounts, projects and prompts, ClickHouse for traces and scores, Redis for the job queue and cache, an object store for raw events, media and exports, and two application containers that have to agree on the same secrets. The published minimums add up to roughly eleven virtual cores and twenty-five gigabytes of memory before you have sent a single trace.

The failure that catches people is quieter than a crash. ClickHouse ships its own system log tables without a retention rule, so an instance with no traffic at all still fills its disk. When you notice and delete old traces, the space does not come back: lightweight deletes mark rows rather than removing them, and the default merge ceiling means the largest parts never merge, so the bytes stay where they are. The forum threads on this run long.

REVIEWS

Hear from customers ​like you​​​​​​​

Successful businesses and professionals around the world rely on DANIAN every day

USE CASES

Three teams who run Langfuse on DANIAN

Three teams, composited from setups we see rather than named accounts. What they share is a per-seat bill that stopped making sense and a stack nobody wanted to operate. What differs is the region, the retention window and what else runs on the account.

6-PERSON AI PRODUCT STUDIO

Moved off per-seat tracing after the fourth hire

Instance in an Asia-Pacific region, one project per client environment, staging traces on a short retention window and production on ninety days. They run n8n on the same account for the nightly export job. Six people, one nine-euro line.

12-PERSON FINTECH DATA TEAM

Retention set by their record schedule, not a vendor default

North American region chosen so trace data stays in one jurisdiction. Scheduled exports land in a bucket they own rather than ours. The encryption key sits in their password manager as well as in our backup set.

4-PERSON RESEARCH GROUP

Three services, one trace

Instance in a Europe region. Model calls come from a Python service and retrieval from a Go service, both emitting OpenTelemetry spans to the same endpoint, so one trace covers the whole pipeline. No seat arithmetic when a student joins for a term.

COMPARISON

Four ways to run Langfuse

Four realistic ways to get Langfuse running, priced at one, five and ten people. The hosted product meters seats and traces. Self-hosting costs no licence fee and a lot of attention. The home server is cheap until the disk fills. Ours is nine euro.

Path1 user5 users10 usersWhat you operate
Proprietary SaaS (LangSmith)~$989/mo~$1,145/mo~$1,340/moVendor runs it. Seats and trace volume both meter. Self-hosting is enterprise-tier only.
Self-host on a VPS$72/mo$72/mo$72/moYou run five services, the schema migrations, the disk and the upgrades.
Home server~€900 once + ~€12/mo powersamesameYou also own the uptime, the backups, the bandwidth and the 03:00 alert.
DANIAN managed Langfuse€9/mo€9/mo€9/moWe run it. You choose the region and keep a dump on request.

Cost is not the only axis. The hosted path caps how long traces live unless you pay for the longer window, and the data sits in the vendor's account. The two self-hosted paths hand you the retention rule and the location and charge you in attention instead. We take the attention and leave you the control.

BY INDUSTRY

Langfuse for specific industries

Four sectors put unusual demands on an observability tool. They need the trace record to outlive the supplier contract, and they need to say where it is stored. Each block names the rule that drives it, what we set differently, and one number that matters.

Broker-dealers keep electronic communications for at least three years under SEC Rule 17a-4, with the first two years readily accessible, and account records for six. An assistant that drafts client-facing research or answers suitability questions produces records inside that perimeter from the day it ships. We provision Langfuse in the region you nominate, set the trace retention window to match your record schedule rather than a vendor default, and hold the encryption key and salt in your backup set so an archived instance can still be opened years later. The workflow we see most: every prompt and completion from the advice assistant is traced with the client reference in metadata, scored by a policy evaluator, and exported on a schedule to a long-term store in a bucket you own rather than one of ours.

Records showing how protected health information was handled must be kept six years under 45 CFR 164.316(b)(2)(i), which is longer than most observability tools keep anything by default. A triage or clinical-documentation assistant produces exactly that kind of record on every call. We run your instance in the region you choose and set the retention window on the trace tables to your six-year figure instead of the thirty-day default that ships with most hosted products. The pattern that repeats: identifiers are swapped for a study reference inside the SDK before the span ever leaves your application, so the mapping never enters the observability stack, while the trace still carries enough structure to reconstruct the decision path.

Regulation 2024/1689 requires providers of high-risk systems to build in automatic event logging under Article 12, and to keep the technical documentation for ten years after the system is withdrawn under Article 18. Ten years is longer than most supplier contracts run. Langfuse gives you the logging surface; running it on infrastructure you control gives you the ten-year part. We provision in the region you nominate, pin the retention window to the lifetime of the system rather than a rolling window, and hand over a full dump on request so the record survives us as a supplier. A common workflow: each decision is traced with the input, the model identifier, the evaluator score and the human override if there was one.

The standard here is not a statute but the OpenTelemetry semantic conventions for generative AI, which is what lets one agent trace hold spans from your orchestration layer, your retrieval service and your model calls together. Teams shipping agents hit the same wall: a single agent turn is not one model call, it is fifteen, and per-unit hosted pricing multiplies accordingly. We provision with the OTLP endpoint open so services already emitting spans need no rewrite, tune the queue and deletion settings for production trace volume rather than demo volume, and leave sign-up closed so the instance stays inside your team. One project per customer environment is the usual layout.

FAQ

Frequently asked questions

Three groups. The first is about running Langfuse itself, which is where most of the surprises live. The second covers moving an existing instance across to us. The third is about DANIAN as a platform and applies the same way to every app in the catalogue.

01

Technical and configuration

We run the MIT-licensed core only: tracing, prompt management, evaluations, datasets, the playground, annotation queues, dashboards and single sign-on. The enterprise licence key stays unset, so no module requiring a paid agreement with the upstream vendor is ever loaded on your instance.
A production install is five moving parts. Postgres holds accounts, projects and prompts. ClickHouse holds traces and scores. Redis carries the job queue and cache. Object storage keeps raw events, media and exports. Web and worker run as separate containers.
ClickHouse ships its own system log tables without a retention rule, so they grow whether or not anyone sends data. We set time-to-live values on those tables when the instance is provisioned. It is the most common self-hosting surprise, and it is removed before you see it.
Nothing, because we do not regenerate them casually. The encryption key and salt are generated once, held identical across the web and worker containers, and carried in the backup set. Losing either is what turns a routine upgrade into an instance that cannot decrypt its own keys.
Yes. We switch off open registration once your owner account exists, so the login page stops accepting new self-service accounts and everyone after that is invited from inside the project. If you would rather keep it open for an internal team, say so and we leave it on.
Yes. Langfuse ingests OpenTelemetry directly, so anything already emitting OTLP spans can point at your instance without a rewrite. The SDK itself is a thin layer over the standard client, which is why mixed instrumentation from several services lands in one trace.

02

Migration and onboarding

21 datacenter locations across six continents. You choose the region at provisioning. Application data sits in the region you choose; pick whichever is closest to your users or matches your data-residency preference.
Yes. Request a region migration from the dashboard and we run the move in the background. The system emails you when the migration completes; total transfer time depends on data volume but typical instances finish in a few hours. There is no extra charge for a region change.
Yes. Full data export is available at any time, in a portable format you can bring to any infrastructure.
Yes. You point a record at the instance and we issue and renew the certificate. If you prefer, send us the registrar login and we make the record ourselves. Custom domains are included at the nine-euro price rather than sold as an extra.

03

Billing, support, and platform

€9 covers everything we do for that app: hardware in the region you choose, daily off-site backups with one-click restore, automatic security patches and version upgrades, 24/7 monitoring, SSL and firewall, and engineering support on email and the ticket section in the client portal. There are no setup fees or hidden line items. For more info see our Pricing page.
If you decide to continue, we charge €9/app/month from day 8. If you don't, the trial ends and you can export your data. No card is required for the trial, and we never auto-charge you without explicit consent.
No. The €9/month is flat regardless of how many users log into your app. Add 5 users or 50; the price doesn't change.
24/7 email and ticket support, both staffed by engineers who run the systems. We handle DNS configuration, SMTP setup, app integrations, performance tuning, troubleshooting, and migration help. Response time is typically under an hour. There is no tier system — every customer gets the same support.
Yes. Cancel from the dashboard. We don't charge a cancellation fee, we don't lock data, and we will export your data to you on request before deletion. data to you on request before deletion.
Every customer instance is backed up daily to a separate region from the primary. We test restores. You can request a restore at any backup point within the retention window — usually 7 days for daily backups.
Your application data sits in the region you choose at provisioning — 21 datacenter locations across six continents. Account-level data (billing, account email, support ticket history) is processed centrally. Application data region is picked by you, per app.
99.9% uptime SLA on every app, every tenant. Service credits are documented at danian.co/service-level-agreement. The status page is located at status.danian.co.
When your tenant approaches the resource ceiling — the base tier holds 1 vCPU/RAM, 30 GB storage — we notify you. Resource upgrades happen with your explicit consent; we will not upgrade your tenant or charge you without it.
We wait. We don't suspend the app or delete your data on the first failed charge. We email you, you fix the card on file, and we continue.
Invoices can be downloaded from the billing dashboard in PDF the day each charge succeeds. EU VAT is added where applicable and the VAT-reverse-charge regime applies for VAT-registered businesses with a valid number.
150+ open-source apps across automation, team chat, file sync, analytics, AI, password management, email marketing, dev tools, project management, smart home, CMS, and federated social. See the full catalog →
Yes. Every instance comes with a web-based terminal and a file manager in your DANIAN management dashboard. Useful for managing your data and customizations.
Resources scale with your usage. If your app needs more vCPU, RAM, or storage, we add it — and we ask first before any change to your plan. €9 is the floor; resource-heavy workloads may price higher, but you'll always know in advance.
Yes. We have both a Partner program and an Affiliate program available. Anybody can sign up.
No contract. No minimum commitment. Cancel anytime from the dashboard with one click. The 7-day free trial requires no credit card. After the trial converts to paid, you can still cancel at any month without notice or penalty.

DEPLOY IN YOUR REGION

21 datacenter locations on six continents

Pick the region closest to your users.

United States, Germany, Finland, Singapore, Australia, Brazil, Canada, Netherlands, UK, Spain, Italy, France, Sweden, Malaysia, India, Japan, Mexico, Poland, South Korea, Chile, South Africa and more coming soon

Global Reach Map

Try managed Langfuse for 7 days

No card. Cancel from the dashboard.