LightningDB Free

Your graph. Your machine.

Download Free for Linux and start building. No account, license key, or activation required.

Version 0.2.0rc1 Release candidate

Choose your Linux downloadInstallation guide

Choose your Linux build.

Free includes the local graph engine and Browser, with up to 2 CPU threads. Tested on Ubuntu 22.04 and 24.04; requires glibc 2.35 or later.

Linux downloads are available now. macOS and Windows builds will follow.

Create your first graph.

Extract your download, open the folder, and start the shell. Choose a data location outside the installation folder.

1. Start the shell

./bin/graphdb shell --db "$HOME/lightningdb-data/my-graph.gdb"

A new file starts an empty graph. Add a node and run your first query:

CREATE (c:Customer {name:'Acme'});
MATCH (c:Customer) RETURN c.name;
:quit

2. Open the Browser

./bin/graphdb serve --db "$HOME/lightningdb-data/my-graph.gdb"

Open http://127.0.0.1:7475/ to query your graph in the local Browser.

Your graph stays on your machine. Free runs on CPU; commercial editions are available separately.

Install with pip instead

On a supported Linux machine with Python 3.10 or later, create a virtual environment and install directly from our download directory:

python3 -m venv .venv-lightningdb
. .venv-lightningdb/bin/activate
python -m pip install --no-index \
  --find-links https://lightninggraph.com/downloads/0.2.0rc1/ \
  'lightningdb[engine]==0.2.0rc1'

graphdb shell --db "$HOME/lightningdb-data/my-graph.gdb"

This installs the Free engine and Python client. pip selects the wheel for your Linux architecture. Model weights are not included.

View package files and checksums
Import CSV or a Neo4j dump

Import a directory containing nodes.csv and rels.csv, then query the saved graph:

./bin/graphdb migrate --from /path/to/csv --out "$HOME/lightningdb-data/imported.gdb"
./bin/graphdb query --db "$HOME/lightningdb-data/imported.gdb" \
  --gql 'MATCH (n) RETURN count(*)'

A compatible Neo4j 5.26 dump can be supplied with --from /path/to/neo4j.dump. Dump import also requires Docker and the Neo4j Python driver. The download includes detailed instructions in share/doc/lightningdb/TUTORIAL.md.

This release supports a defined Cypher subset, one label per node, and relationships without properties. Unsupported inputs are rejected. Test your workload before production use.

Planning a commercial deployment?

Talk with us about your workload, deployment, and support requirements.

Talk to the teamView plans