Clone a Demo Database

Clone a pre-populated demo database from the public templates server to your local TerminusDB instance. No account required — just a running TerminusDB on localhost:6363.

Available templates

The public templates server at data.terminusdb.org hosts curated datasets you can clone in one command:

Template	Documents	Use case
Star Wars	~85 (People, Films, Planets, Species, Starships, Vehicles)	Learning WOQL queries and GraphQL, following tutorials
Ecommerce	~155 (Customers, Orders, OrderLines, Products, Categories)	Business scenario tutorials (branch, diff, merge)
Nuclear	Power plants, reactors, elements, countries	Graph data with nested structure and enums
Lego	Sets, parts, minifigures, themes, colours	A larger dataset (~60 MB) for real-world scale
Sandbox	A handful of products	The live target for branch/diff/merge docs examples
MyDB (tdb-example-mydb)	Project tracker	Rich commit history — 11 commits, branches, merges

Clone Star Wars

curl -u admin:root -X POST http://localhost:6363/api/clone/admin/star-wars \
  -H "Content-Type: application/json" \
  -H "Authorization-Remote: Basic cHVibGljOnB1YmxpYw==" \
  -d '{"remote_url": "https://data.terminusdb.org/public/star-wars", "label": "Star Wars", "comment": "Star Wars demo dataset"}'

Then explore it: Explore a Real Dataset

Clone Ecommerce

curl -u admin:root -X POST http://localhost:6363/api/clone/admin/ecommerce \
  -H "Content-Type: application/json" \
  -H "Authorization-Remote: Basic cHVibGljOnB1YmxpYw==" \
  -d '{"remote_url": "https://data.terminusdb.org/public/ecommerce", "label": "Ecommerce", "comment": "Ecommerce tutorial dataset"}'

Then explore it: Explore an Ecommerce Dataset

How it works

The clone operation uses TerminusDB's git-for-data architecture:

Your local instance contacts data.terminusdb.org (authenticated with public:public read-only credentials)
The entire database — schema, documents, and commit history — is pulled to your local instance
The clone is now yours: query it, branch it, modify it. Changes stay local.

Cloning is efficient even for large datasets because TerminusDB uses content-addressed storage — only unique data layers are transferred.

Semantic search on a clone

Every template ships with embedding templates baked into its schema. If your local TerminusDB has vector search enabled, the clone indexes itself — the search indexer is notified automatically once the clone finishes, and again after every pull that brings in new commits. No manual indexing step.

You need two things locally:

TerminusDB with the vectorlink indexer enabled (TERMINUSDB_INDEXER_BACKEND=vectorlink)
an embedding endpoint — a local Ollama instance on port 11434, or any OpenAI-compatible embeddings API. The vectorsearch docker compose stack ships both; see VectorLink Quickstart.

Once the indexer has worked through the documents, search the clone:

curl -u admin:root 'http://localhost:6363/api/search/admin/star-wars?q=jedi+knight'

The public server itself is clone-only — it does not index the products. Embedding always happens on your own instance, with your own embedding service.

Next steps

Get Started — the complete quickstart tutorial
Explore a Real Dataset — Star Wars dataset walkthrough
Explore an Ecommerce Dataset — business scenario walkthrough
WOQL Basics — learn the query language
GraphQL Basics — query with GraphQL