The Open-Source Software Behind AIblogly
In short: AIblogly is built almost entirely from open-source software - code anyone can read, use and change under its licence. The site runs on Next.js and React, stores its data in Postgres through Prisma, and is styled with Tailwind CSS. This page lists that software, with links and licences. Many of the AI models the site covers are open-weight, a separate idea explained in open-weight AI models explained.
What is open-source software?
Open-source software is software whose source code is published under a licence that lets anyone read, use, change and share it. The licence sets the conditions. Permissive licences such as MIT and Apache 2.0 allow almost any use, including commercial use, as long as the licence notice is kept. Most of the web, including this site, is built from open-source projects maintained in public on GitHub.
What open-source software runs AIblogly?
Everything below is read from the project's package.json, the file that lists what the site installs. Versions are the ones installed at the time of writing.
Framework
| Project | Version | What it does here |
|---|---|---|
| Next.js | 16.3.8 | The web framework: page routing, server rendering and caching |
| React and React DOM | 19.2.8 | Builds the interface from components |
| TypeScript | 5.2.2 | JavaScript with types, so mistakes are caught before the code runs |
Data
| Project | Version | What it does here |
|---|---|---|
| Prisma | 6.7.0 | Reads and writes the Postgres database - every model, guide and news story is a row |
| PGlite | 0.5 | Runs Postgres locally for development, without a server install |
| hyparquet | 1.30 | Reads Parquet files, the format of LMArena's published leaderboard dataset |
Interface
| Project | Version | What it does here |
|---|---|---|
| Tailwind CSS | 3.3.3 | Styling, with @tailwindcss/typography for article text |
Lucide (lucide-react) | 0.446.0 | Icons |
Motion (framer-motion) | 13.1.0 | Interface animation |
| GSAP | 3.15 | Scroll-linked animation on the homepage |
| Recharts | 2.15.3 | Charts on the comparison and benchmark pages |
| Sonner | 2.0.8 | Pop-up notifications |
Content and utilities
| Project | Version | What it does here |
|---|---|---|
| react-markdown | 9.0.1 | Turns guide and news text written in Markdown into web pages |
| remark-gfm | 4.0.0 | Adds GitHub-flavoured Markdown, such as the tables on this page |
| bcrypt.js | 2.4.3 | Password hashing for the admin login |
| jsonwebtoken | 9.0.2 | Signed session tokens |
| Autoprefixer | 10.4.15 | Adds browser-specific CSS prefixes |
One exception to "open source": GSAP is free to use but published under its own licence rather than an open-source one. The particle animation on the homepage is hand-written canvas code, so no library appears for it.
What about open-weight AI models?
Open-weight models are a different thing from open-source software: the trained model file is published, but its training code and data often are not. Many of the AI models this site covers are open weight. What that means, which models publish their weights, and the licences they use are in open-weight AI models explained.
Words to know
- Open source - software whose code is published under a licence allowing use, change and sharing.
- Licence - the legal terms that say how code or a model may be used.
- MIT / Apache 2.0 - two widely used permissive licences that allow commercial use.
- Dependency - an outside software package a project installs and uses.
- Open weights - a model's trained files published for download.
- Custom licence - a licence written for one product, with its own conditions.
Related guides
Frequently asked questions
What is AIblogly built with?+
AIblogly is built with Next.js and React, uses Prisma to read and write a Postgres database, and is styled with Tailwind CSS. Charts use Recharts, animation uses GSAP and Motion, and guide text is written in Markdown and rendered with react-markdown.
What is the difference between open source and open weights?+
Open-source software publishes its source code under a licence that allows use, change and sharing. An open-weight AI model publishes its trained weights for download, but may not publish its training code or data, and its licence may restrict how it is used.
Is Next.js open source?+
Yes. Next.js is developed in public on GitHub under the MIT licence, as is React; Prisma uses the Apache 2.0 licence.
Enjoyed this? Get more in your inbox.
Up to ten of the week's AI news stories, by email once a week.