Aug. 15, 2023, 2:14 p.m. Simon Willison
Datasette Cloud is the new SaaS hosting platform for the Datasette open source project. It enables teams to create a private collaboration space, upload and share data securely with each other, and then selectively publish that data to the world.
Datasette was born out of data journalism, and the initial intended audience for Datasette Cloud is newsrooms: we want to help journalists share data with each other and with the public without needing to configure and run Datasette on their own hosting provider.
We expect that tools which help journalists find stories in data will be of value to all kinds of other companies and organizations as well though, so everyone else is invited to try out the product too!
Teams can create a private space to collaborate on data, hosted in a region close to them. Spaces each run in their own secure container (powered by Fly.io) and get their own
The space administrator can then invite further members to join their space.
Team members can import data into their space in multiple different ways:
- Upload CSV files to create tables
- Load CSV files from a URL
- Import data from government open data portals that are powered by Socrata
- Create new databases by uploading a SQLite database file
- Import data using the Datasette Write JSON API
Once imported, data can be edited inline through the Datasette Cloud web interface. See Introducing datasette-write-ui: a Datasette plugin for editing, inserting, and deleting rows.
Data in Datasette Cloud can then be accessed using all of the tools provided by Datasette:
- The Datasette web interface, with filtering, faceting and the ability to run SQL queries and bookmark the results. See Exploring a database with Datasette for a tutorial and live demos, and Learn SQL with Datasette for information on using SQL to query your data.
- A powerful JSON API, protected by finely-grained API tokens, for accessing tables and executing SQL queries.
- A GraphQL API to all of your tabular data, powered by the datasette-graphql plugin.
All data in Datasette Cloud is stored securely in Fly volumes, and also backed up to S3 using Litestream.
What’s coming next?
We’re working on a number of features that will be coming to Datasette Cloud soon:
- AI-assisted queries. We’ve been researching ways Large Language Models (the technology behind OpenAI’s ChatGPT, Anthropic’s Claude and Google’s Bard) can be responsibly applied to help journalists find stories in data. Read more at llm.datasette.io.
- The ability to publish selected tables and queries to the outside world.
- Data annotations: your team will be able to add comments to columns and rows in your data, making it easier to collaborate on data analysis and find stories in your data.
- Pricing! We want you to be able to pay for the product, so you can feel confident in its sustainable future and we can get paid to keep working on it.
Sign up for Datasette Cloud
Book a demo session over Zoom and be one of the first to try out Datasette Cloud.
Or skip the demo and request preview access.