Quickstart

Graphora offers three ways to get started, from zero-install to full local development. Pick the one that fits your situation.

Option 1: Try the Demo (30 seconds)

No signup, no install. Visit demo.graphora.io, upload a document, and see Graphora extract a knowledge graph in real time. This is the fastest way to see what Graphora does.

Option 2: Google Colab Notebook (2 minutes)

Run the interactive quickstart notebook in your browser. It walks you through defining an ontology, extracting entities from a sample document, and visualizing the resulting graph — all without installing anything locally.

Open in Google Colab

Launch the quickstart notebook

The notebook uses pip install graphora[cli] and walks you through configuring your Gemini API key. You’ll need a Google account and a free Gemini API key.

Option 3: CLI (5 minutes)

Install the Graphora CLI and extract knowledge graphs from the command line. The CLI includes an embedded mode that runs the API locally, so you do not need a separate server.

Install the CLI

pip install graphora[cli]

This installs both the Python client library and the graphora CLI command.

Configure your Gemini API key

Graphora uses Google Gemini for AI-powered extraction. Get a free API key from Google AI Studio, then configure it:

graphora config init --api-key "YOUR_GEMINI_API_KEY"

You can also set it as an environment variable: export GEMINI_API_KEY="YOUR_KEY"

Extract a knowledge graph

graphora extract document.pdf --output graph.json

The CLI will process the document and write the extracted graph to graph.json. Schema is inferred automatically from your document content.

Explore the output

The output is a JSON file containing nodes (entities) and edges (relationships) extracted from your document. You can load it into any graph visualization tool or use it in your application.

# Quick peek at what was extracted
python -c "
import json
graph = json.load(open('graph.json'))
print(f'Nodes: {len(graph[\"nodes\"])}')
print(f'Edges: {len(graph[\"edges\"])}')
for node in graph['nodes'][:5]:
    print(f'  {node[\"labels\"]}: {node[\"properties\"].get(\"name\", node[\"id\"])}')
"

For more CLI options, see the CLI Reference.

Next Steps

Once you have seen Graphora in action, you can dive deeper:

Platform User Guide

Complete walkthrough of the web UI workflow

Schema Copilot

Build ontology schemas through natural language chat

Local Development

Run Graphora locally with zero-config in-memory mode

Python Client Library

Integrate Graphora into your data pipelines programmatically

​Quickstart

​Option 1: Try the Demo (30 seconds)

​Option 2: Google Colab Notebook (2 minutes)