Fully Managed Laminar
as a Service

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

Laminar is an open-source observability platform built for AI agents. It records the full execution tree of an agent run, stores the spans in a column store, and layers evaluations, datasets and browser session replay on the same data.

✓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

Laminar

Laminar

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 →
Laminar Preview Image

ABOUT THE SOFTWARE

What is Laminar

Laminar is an open-source observability platform for AI agents. It captures each run as a tree of nested spans covering model calls, tool calls, retries and browser steps, then adds evaluations, datasets and dashboards on top of the same stored data.

The project is maintained by LMNR AI, Inc., a Y Combinator company, and published under the Apache License 2.0. The codebase is a Rust ingest server, a Next.js interface, and client libraries for Python and TypeScript. Ingest is OpenTelemetry-native, so anything that already speaks OTLP can send spans without a proprietary agent in the path.

A working deployment is not one container. The lightweight stack runs the interface, the Rust application server, PostgreSQL for metadata, ClickHouse for span data and Quickwit for full-text search. The production stack adds a message queue, a cache and a separate consumer process. Teams that publish their use of it include Browser Use, OpenHands and Rye.

FEATURES

What Laminar does

Laminar records what an agent actually did, not just what it finally answered. Each run becomes a trace of nested spans, and evaluations, datasets and dashboards read from that same store, so debugging and measurement never drift apart into two systems.

Agent trace tree

Every run is captured as nested spans: the planning step, each model call, each tool call, retries and failures, with inputs and outputs attached at each level.

OpenTelemetry-native ingest

Spans arrive over OTLP using the GenAI semantic conventions. Existing collectors and exporters keep working, and there is no closed agent sitting in the request path.

Evaluations

Run scored evaluations against saved inputs, offline or as production traffic arrives, and compare scores across prompt versions and model choices in the same interface.

Datasets and labelling

Promote interesting traces into datasets, label them by hand or with a scoring model, and reuse them as the fixed input set for the next evaluation run.

Browser session replay

For agents that drive a browser, the session is recorded alongside the trace, so a failed step can be watched rather than reconstructed from log lines.

SQL editor and dashboards

Query span data directly with SQL and pin the results to dashboards, which is possible because the underlying store is a column database rather than a closed index.

Python and TypeScript SDKs

Client libraries with automatic instrumentation for common model providers and agent frameworks, so a first trace usually needs a decorator and an environment variable.

Redaction service

A separate service strips sensitive values out of span payloads before they are stored, which matters when prompts carry customer records.

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 Laminar 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 Laminar

Hosted tracing platforms bill by the trace, and that meter has been repriced upward while retention windows got shorter. A team that instrumented one agent last year now watches a bill scale with traffic instead of headcount. Running Laminar yourself fixes the price and keeps the spans in your own store.

The part that catches people out is the column store. Laminar keeps span data in ClickHouse, and ClickHouse reaches object storage over the S3 API. On a cloud host that hands out short-lived credentials from a metadata endpoint, those credentials are rejected, and the fix is to issue static interoperability keys and write them into the storage configuration by hand. Get that wrong and the container restarts in a loop, usually with an XML parse error from a placeholder that was never replaced.

The second trap is one-way. Quickwit records the storage location of an index at the moment the index is created. Boot the stack once against a placeholder bucket and that location is written into the index metadata permanently. Changing the setting afterwards does not move anything, and the recovery is to drop and rebuild the search indexes. There is also an upgrade step where the queue and search stateful sets have to be removed with their pods orphaned before the new version will start.

We set those choices once, at provisioning, and carry them forward across upgrades:

Object storage for span data wired with static interoperability keys, never with credentials borrowed from the host metadata service.

The search index root location written at first boot and pinned, so indexes are never stranded against a placeholder bucket.

A per-instance encryption key generated at provisioning, held in the secret store and carried in the backup set, so stored provider keys survive a restore.

Upstream product telemetry switched off by default, and switched back on only if you ask for it.

A per-project ingest rate limit and queue reconnect settings tuned before the first span arrives, rather than after a burst takes the queue down.

REVIEWS

Hear from customers ​like you​​​​​​​

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

USE CASES

Three teams who run Laminar on DANIAN

Three shapes of team come to managed Laminar: a product studio watching a metered tracing bill outgrow its seat bill, an automation shop that needs one instance per client, and an engineering team that has to keep the records of what a model decided.

6-PERSON AI PRODUCT STUDIO

Moving off metered tracing before the next pricing change

Traces from a support agent were costing more than the seats. They moved to a managed instance in a region close to their inference provider, kept the OTLP exporter they already had, and set the ingest rate limit per project so a runaway retry loop cannot flood the queue.

4-PERSON AGENT AUTOMATION SHOP

One instance per client, on one predictable line item

They build browser agents for retail clients and cannot mix client data in one workspace. Each client gets its own instance at EUR 9, in the region that client asks for, with browser session replay switched on and the redaction service in front of span storage.

12-PERSON LENDING PLATFORM TEAM

Keeping the record of what the model was asked and answered

Their bank partner asks for evidence of ongoing monitoring on any model in the decision path. Traces, evaluation scores and the dataset each score was computed against live on one instance, in a single region, with daily off-site copies held under the retention window their partner sets.

COMPARISON

Four ways to run Laminar

Four paths, the same software underneath. A hosted tracing platform charges per seat and then again per trace. A server you rent is cheap until the column store needs attention. A machine at home is cheapest on paper. DANIAN is EUR 9.

Path1 user5 users10 usersWhat you also carry
Proprietary SaaS (LangSmith)USD 39/moUSD 195/moUSD 390/moPer-seat price plus a metered charge per thousand traces, and a retention window that is priced separately from ingest.
Self-host on a rented serverUSD 24/moUSD 24/moUSD 24/moA production-class server, sized up once ClickHouse, search and the queue are all resident. Storage keys, index locations, upgrades and backups are yours.
Home serverUSD 900 onceUSD 900 onceUSD 900 onceAn HP ProLiant ML30 Gen10 or similar, plus electricity. No off-site copy, a residential address on the certificate, and your own uptime.
DANIAN managed LaminarEUR 9/moEUR 9/moEUR 9/moNothing. Region of your choice, daily off-site backups, updates applied, and support from the people running it.

BY INDUSTRY

Laminar for specific industries

Some buyers want Laminar for debugging. Others need it because a supervisor, a regulator or a partner bank asks what the model was given and what it returned. Those four settings put different demands on retention, region and who can read a span.

A sponsor bank applies SR 11-7, the supervisory guidance on model risk management issued as OCC Bulletin 2011-12, to any model in a credit or fraud decision path, including models a vendor supplies. That guidance asks for documentation, independent validation and ongoing monitoring rather than a one-time sign-off. On DANIAN we run the instance in a single named region and hold the evaluation datasets on the same instance as the traces they scored, so a monitoring pack is assembled from one store. A typical workflow is a weekly scored evaluation against a frozen dataset of declined applications, exported for the partner review. Backups are retained daily and held for the window the sponsor sets, commonly three years.


Firms building assistants that touch client communications sit under SEC Rule 17a-4 and FINRA Rule 4511, which require records to be preserved for six years, with the first two years kept readily accessible. Prompts and completions that summarise or draft a client message fall inside that record set. On DANIAN we pin the instance to one region so the record location is a single answer, and put the redaction service in front of span storage so account numbers never reach the column store. A common workflow is exporting a month of traces for the supervisory review file. Six years is the number to plan storage against, not six months.


Regulation 2024/1689 requires a high-risk system to record events automatically over its lifetime under Article 12, and requires the deployer to keep those logs for at least six months under Article 19 unless another rule sets a longer period. Laminar spans are a natural place for that record, provided they are not silently rotated away. On DANIAN the retention window is a choice you make rather than a tier you buy, and the instance runs in the single region you name. A typical workflow is a monthly export of traces for the technical documentation pack. Plan for six months as the floor, not the target.


Teams building assistants that touch study records work under 21 CFR Part 11, where section 11.10(e) requires secure, computer-generated, time-stamped records of operator entries and actions, and 21 CFR 312.62(c) requires investigator records to be kept for two years after a marketing application is approved. On DANIAN we run the redaction service ahead of span storage so participant identifiers never land in a prompt payload, and we keep the instance in one region with daily encrypted off-site copies. A common workflow is attaching the trace and its evaluation score to the change record for a prompt revision. Two years past approval is the retention floor to size for.


FAQ

Frequently asked questions

Three groups of questions. What the software does and how we run it, what moving an existing setup across involves, and how billing, regions and support work on the platform underneath. Anything not answered here gets a reply from an engineer, not a form.

01

Technical and configuration

The interface, the Rust application server, PostgreSQL for metadata, ClickHouse for span data and Quickwit for full-text search. Production instances also run a message queue, a cache and a separate consumer process. All of it sits inside one instance on one EUR 9 line item.
Set the base URL to your own domain and keep the standard ports for HTTP and gRPC ingest. The Python and TypeScript libraries take the same three settings they take against any deployment, so example code from upstream works without edits.
We issue static interoperability keys and write them into the storage configuration directly. ClickHouse speaks the S3 API, and short-lived credentials handed out by a host metadata service are rejected by it, which is the single most common reason a self-hosted instance restarts in a loop.
We set the index root location once, at first boot, and pin it. Quickwit records that location in the index metadata when the index is created, so a first start against the wrong bucket is permanent and only a full index rebuild recovers it.
We generate a per-instance encryption key at provisioning and keep it in the instance secret store. It encrypts the model and project keys you save. The key travels in the backup set, so a restore does not leave you with credentials that can no longer be read.
We set a per-project ingest rate limit and tune queue reconnect and insert timing before the first span arrives. A retry storm inside an agent is the usual cause, and the limit keeps one noisy project from taking ingest down for the others.
The project collects anonymous product telemetry by default. We switch that off when we provision, and we switch it back on only if you ask. Nothing about your spans, prompts or model keys is sent anywhere by us.
GitHub, Google, Azure AD, Okta and Keycloak are supported by the application. We wire up the one you use and put your own domain and certificate in front of it, which is included rather than a paid extra.
Almost. Tracing, search, dashboards, the SQL editor, evaluations, datasets, labelling and browser session replay are all in the open build. The failure clustering feature and the email and chat alerting need a licence key sold by the project itself, which we do not resell.

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, for new traffic. You change the base URL in your exporter and spans start landing on your instance. Historical traces held by another vendor generally cannot be transferred, so most teams run both for a short overlap and then stop paying the meter.
It should. Ingest is OTLP and follows the GenAI semantic conventions, so a collector you already run can add an exporter pointed at your instance. Teams already emitting spans from other services often send both to the same place.
Yes. Send us a PostgreSQL dump and the span data from your column store and we load both onto the new instance. We check that saved provider keys decrypt after the load, because that is the step that usually breaks a hand-run migration.
Yes. Custom domains are included on every app. We issue and renew the certificate, and we can walk you through the DNS records if that is not something you deal with often.
You get dumps of both datastores and take them wherever you want. The software is Apache-licensed and self-hostable, so there is no format or key that only we can read. Cancelling is a button in the dashboard.
Yes. Send us the provider you use and we configure it on your instance. This is part of support rather than an onboarding fee, and it is included at the same EUR 9.

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 Laminar for 7 days

No card. Cancel from the dashboard.