Skip to main content
Decentralized Deployments

Search Without Clusters

In the Cloud. On-Prem. Air-Gapped. At the Edge.

Search should be invisible. It should just work.

The cluster was never the point — it was the price of admission. Sizing it, sharding it, rebalancing it, and keeping a quorum alive became a discipline of its own, staffed by search administrators and DBAs whose job is keeping the search engine healthy rather than using it. Lucenia removes the cluster entirely. Nodes are peers on a mesh: they join, they serve, they leave, and the system reconciles without a coordinator and without you.

Lucenia is built differently.

The Reality

The Cluster Is the Part You Actually Maintain

Not the search. The cluster.

Someone has to size the cluster before anyone knows the workload

Shard counts are chosen up front and expensive to change later

Rebalancing, hot shards, and split-brain become a full-time job

Capacity is provisioned for peak and paid for at idle

Every upgrade is a coordinated, cluster-wide event

Most search platforms today:

  • Require a cluster to be planned, sized, and tuned before the first query
  • Assume dedicated search administrators and DBAs keep it healthy
  • Degrade when a node leaves, and stay degraded until a human intervenes
  • Turn growth into a migration project instead of a setting

So you staff a team to keep the search engine alive, instead of a team that uses it.

The Mechanism

Disposable Primaries

In a cluster, losing the wrong node is an incident. Here it is a Tuesday.

claim epoch7

Steady state. One node holds the claim at epoch 7 — every other node is an equal peer.

No consensus protocol. Ownership is a claim with an epoch, and the highest epoch wins. There is no Raft group to keep healthy and no quorum to lose.
Ownership is not durability. Data lives in content-addressed segments and a write-ahead log, so the node holding the claim can vanish without taking anything with it.
Churn is the normal case. Nodes joining and leaving is the design point, not the failure path — which is what makes edge and intermittently-connected sites viable.
Your Reality

Your Data Lives Everywhere. Retrieval Should Too.

And it is not going to consolidate into one cluster to suit your search vendor.

On-Prem Data Centers

ERP systems, file shares, manufacturing systems, internal apps

Air-Gapped & Restricted Networks

Defense, healthcare, regulated research, critical infrastructure

Hybrid Architectures

Legacy systems + private clouds + modern SaaS

Public Cloud

Modern applications, analytics, AI workloads

A mesh spans all of this. A cluster asks you to pick one and migrate the rest.

Lucenia's Approach

Every Node Is a Peer, Wherever It Runs

The same engine, the same behaviour, and nothing to coordinate — on-prem, air-gapped, hybrid, or at the edge.

Fully On-Prem

  • Add a node; it joins the mesh and starts serving
  • No coordinator to stand up, no quorum to protect
  • Full control over updates and versions

Ideal for regulated industries, manufacturing, and traditional enterprise IT

Air-Gapped

  • No internet, no control plane, no phone home
  • Peers reconcile directly with each other
  • Offline licensing and updates

Ideal for defense, healthcare, critical infrastructure

Hybrid

  • One mesh spanning on-prem and cloud
  • Data stays where policy says it stays
  • Consistent API and experience everywhere

Ideal for M&A integration, digital transformation, and multi-cloud strategies

Edge

  • Runs where the data is created, not where a cluster lives
  • Survives churn — nodes come and go by design
  • Scales to zero when nothing is asking

Ideal for field systems, disconnected sites, and constrained hardware

Enterprise-Grade

What You Stop Doing

The clearest measure of invisible infrastructure is the work that disappears.

YOU DROP

No Cluster to Operate

  • Capacity planning
  • Shard sizing
  • Rebalancing
  • Split-brain drills

There is no cluster state to protect, so there is no cluster state to lose.

YOU DON'T STAFF

No Specialists to Hire

  • Search administrators
  • Database administrators
  • Tuning specialists
  • Cluster on-call rotations

Retrieval should not require a team whose job is keeping retrieval alive.

YOU GET

Same Everywhere

  • Same APIs
  • Same features
  • Same performance
  • Same support

No feature degradation based on where you deploy.

Built For You

Built for Teams Who Would Rather Not Think About It

Lucenia is built for organizations that:

  • Do not want a search team to have a search platform
  • Cannot "just move everything to the cloud"
  • Operate under strict regulatory frameworks
  • Run mission-critical systems across sites that come and go
  • Want retrieval to be infrastructure, not a project

Clusterless isn't a deployment option. It's the architecture.

Stop Operating a Cluster to Get a Search Result

You shouldn't have to sacrifice:

SimplicityforScale
SovereigntyforFeatures
ControlforConvenience

With Lucenia, you don't.

No Clusters. No Admins. Every Environment.

Lucenia runs as a decentralized mesh exactly where you need it — on-prem, air-gapped, hybrid, or at the edge — with no cluster to plan and no specialists to staff.

Try locally in one minute

curl -sSL https://get.lucenia.dev | bash
Reference Guide
OR

Deploy for production

Start Free Trial

Or, deploy on-prem

Retrieval that behaves like infrastructure: present, dependable, and out of your way.