TETRA Embedded · On-prem

Build Tetra into your product.

One small binary you ship inside your application: on a laptop, at the edge, or in your customer’s private cloud. For Tetra, on-prem and embedded are the same thing — there is no separate cluster to run, so there is nothing to install in the datacentre but your own product.

  • Embeds as a single file, so you distribute it like any other library.
  • Drop-in replacement. Keep your current queries and connections.
  • Runs anywhere, air-gapped and offline-capable, with no tuning to maintain.
  • Lost your embedded graph database? Cypher-compatible, so the port is a short one.
your-product
# one file. ships with your app.
$ ./tetra serve graph.tetra
 
# same driver, unchanged connection
DB_URL = "bolt://localhost:7687"
66MBIdle memory
1File to deploy
0Tuning required

Where it runs

Four shapes, one binary.

Inside your app

One file, linked in and distributed like any other library. No sidecar service, no install step for your customer.

At the edge

Runs on a box in the field with no cloud to reach. Local, low-latency graph queries where the data actually is.

Air-gapped

Offline-capable by construction — no phone-home, no license server, nothing that needs a network to answer a query.

Private cloud

Ships into the customer’s own environment for the deals where data cannot leave it. On-prem and embedded are one thing here.

Coming off an embedded graph?

A supported home, not a fork.

If the embedded, Cypher-speaking engine you built on was discontinued, Tetra is a like-for-like replacement you can actually depend on: maintained, single-binary, and Cypher-compatible, so the migration is queries and drivers rather than a re-architecture.

See the use cases
The move, in short

Repoint the Bolt driver, import the data, verify. Your openCypher queries run unchanged.

Single binary you ship and control — nothing to operate, nothing to babysit.