🪐 Cosmograph

A powerful JavaScript graph visualization library built around a GPU-accelerated force layout engine with 2D and 3D support. It is especially suited to knowledge graphs—large networks of entities, relationships, and ontologies you need to see and navigate in the browser—as well as semantic maps from machine learning embeddings. Cosmograph keeps those graphs responsive at scale and pairs the renderer with interactive tools for search, filtering, and exploration.

✨ Features

• 🕸️ Knowledge-graph friendly: visualize entity–relationship networks at scale, with layouts and interactions that help you spot structure, neighborhoods, and outliers
• 🧊 2D and 3D: switch between a flat view and a GPU 3D force layout with an orbit camera (spaceDimensions: 3), including optional Z coordinates for embeddings
• 🧠 High-performance processing of millions of points and edges on modern hardware
• 🧩 Ready-to-use interactive components including a timeline, search, histograms, legends, and more.
• 🛠️ Flexible configuration API for full customization over graph behavior and appearance
• 🎨 Rich set of automatic data-based coloring and sizing strategies for points and links
• ✨ Real-time addition, removal, and positioning of points and links
• 📊 Support for standard data formats including .csv, .parquet, .arrow and .json
• 🔍 SQL capabilities for powerful data transformations and filtering
• 🔗 Direct connection to tables in external DuckDB instances
• ⌨️ TypeScript support featuring comprehensive types and API hints

🚀 Quick Start

1️⃣ Install the package:

npm install @cosmograph/cosmograph

2️⃣ Get the data, configure the graph and render Cosmograph in the provided container:

import { Cosmograph, prepareCosmographData } from '@cosmograph/cosmograph'

// Points and links data can be:
// - Array of objects
// - File (.csv/.tsv, .parquet/.pq, .arrow, .json)
// - URL string to a file
// - Apache Arrow Table (binary data Uint8Array/ArrayBuffer)
// - DuckDB table name if connection is passed into the Cosmograph constructor

const rawPoints = [{ id: 'a' }, { id: 'b' }, { id: 'c' }]

const rawLinks = [
  { source: 'a', target: 'b' },
  { source: 'b', target: 'c' },
  { source: 'c', target: 'a' },
]

// Create a config to map your data into Cosmograph's internal format
const dataConfig = {
  points: {
    pointIdBy: 'id',
  },
  links: {
    linkSourceBy: 'source',
    linkTargetsBy: ['target'],
  },
}

const createCosmograph = async (container: HTMLElement): Promise<void> => {
  // Prepare data and config for Cosmograph
  const result = await prepareCosmographData(dataConfig, rawPoints, rawLinks)

  // Create Cosmograph instance from prepared data and config
  if (result) {
    const { points, links, cosmographConfig } = result
    const cosmograph = new Cosmograph(container, { points, links, ...cosmographConfig })
  }
}

// Render Cosmograph in the provided container
createCosmograph(document.getElementById('cosmograph-container'))

Check out the Cosmograph documentation for more information and examples.

🛸 Issues and Contributing

While currently closed source, we welcome your feedback! Help improve Cosmograph by submitting bug reports and feature ideas in our issues repository.

💫 License

Cosmograph is licensed under the CC-BY-NC-4.0 license and free for any non-commercial usage. If you want to use it in a commercial project, please reach out to us.

🌌 Cosmograph Universe

@cosmograph/react - Cosmograph for React applications
py_cosmograph - Jupyter widget for data science workflows in Python

👩🏻‍🚀 Contacts

🌎 cosmograph.app
📩 [email protected]
👾 Cosmograph Discord channel