Grist — Modern Relational Spreadsheet You Can Self-Host

Introduction

Grist is a modern, open-source spreadsheet that stores data in proper relational tables instead of flat grids. It gives spreadsheet users the power of a database with familiar formulas, while giving developers a real API and Python-based computed columns. You self-host it and own all your data.

What Grist Does

• Organizes data in relational tables with typed columns and foreign-key references
• Evaluates formulas written in Python instead of traditional spreadsheet syntax
• Provides multiple views (grid, card, chart, calendar) over the same underlying data
• Exposes a REST API for programmatic reads, writes, and webhooks
• Supports granular access rules to control who sees or edits which rows and columns

Architecture Overview

Grist is a Node.js server that manages SQLite-backed documents. Each document is an independent SQLite file, making backups and portability straightforward. The Python formula engine runs in a sandboxed subprocess (gvisor or pyodide) that evaluates column formulas on each change. The frontend is a TypeScript single-page app using GrainJS for reactive DOM updates.

Self-Hosting & Configuration

• Run with Docker using the official gristlabs/grist image with a persistent volume
• Set GRIST_SESSION_SECRET for session security and APP_HOME_URL for your domain
• Configure authentication via SAML, OpenID Connect, or a header-based auth proxy
• Enable sandboxing with GRIST_SANDBOX_FLAVOR=gvisor for untrusted formula execution
• Optionally connect to an external PostgreSQL or S3 for document metadata and storage

Key Features

• Python formulas with full standard library access inside a security sandbox
• Access control rules that restrict data visibility at the row and column level
• Incremental imports that update existing records by matching on key columns
• Webhooks that fire on row changes for integrating with external services
• Document snapshots for point-in-time restore and version history

Comparison with Similar Tools

• Airtable — SaaS with vendor lock-in; Grist is self-hosted with portable SQLite files
• NocoDB — wraps an existing SQL database; Grist manages its own data in SQLite docs
• Baserow — similar concept but uses Django and PostgreSQL; Grist offers Python formulas in cells
• Google Sheets — collaborative but cloud-only; Grist provides local ownership and relational modeling

FAQ

Q: Can I use Grist as a backend for my app? A: Yes. The REST API supports full CRUD and filtering, making it a lightweight backend for small applications.

Q: How do Python formulas work? A: Each computed column has a Python expression. When dependent data changes, Grist re-evaluates affected formulas in a sandboxed Python process.

Q: Is there a row or data limit? A: Limits depend on your hardware. Since each document is an SQLite file, practical limits are in the hundreds of thousands of rows per document.

Q: Can I import from Excel or CSV? A: Yes. Grist supports importing Excel, CSV, JSON, and Google Sheets with column type detection and incremental update matching.

Sources

• https://github.com/gristlabs/grist-core
• https://support.getgrist.com/