Skip to main content
Before contributing to AI Stats, it’s important to understand the layout of the repository.
The project is designed for clarity and scalability - separating data, documentation, and app logic so you can easily find what you need.

​
Overview

Here’s what the root directory of the repository looks like: 
AI-Stats/
+-- apps/
¦   +-- web/ # The main AI Stats Next.js website (contains data)
¦   +-- api/ # The AI Stats Gateway API (Hono/Cloudflare Workers)
¦   +-- docs/ # The Mintlify-powered documentation site
¦
+-- packages/ # SDKs and shared tooling
¦   +-- sdk-ts/
¦   +-- sdk-py/
¦   +-- sdk-go/
¦   +-- sdk-csharp/
¦   +-- sdk-php/
¦   +-- sdk-ruby/
¦   +-- sdk-java/
¦   +-- sdk-rust/
¦   +-- sdk-cpp/
¦   +-- ai-sdk-ai-stats/
¦
+-- scripts/ # Automation scripts and utilities
+-- .changeset/ # Release metadata
+-- turbo.json # Turborepo configuration
+-- package.json # Global scripts and dependencies
+-- README.md # Root readme file

​
Data folders

AI Stats stores its structured data in JSON files under the /apps/web/src/data directory.
These files act as the source of truth for models, benchmarks, and pricing.

​
/apps/web/src/data/models

Contains one JSON file per model.
Each file describes the model’s metadata, pricing, benchmarks, and capabilities. Example:
/apps/web/src/data/models/openai/gpt-5-2025-08-07/model.json

​
/apps/web/src/data/organisations

Describes companies or research groups behind models.
Each JSON file includes name, description, founding date, website, and country. Example:
/apps/web/src/data/organisations/openai.json

​
/apps/web/src/data/benchmarks

Contains benchmark metadata and aggregated results across models.
Each benchmark has an ID, name, description, category, and metrics. Example:
/apps/web/src/data/benchmarks/mmlu.json

​
/apps/web/src/data/api_providers

Lists all supported API providers with their supported models.
This links to both models and organisations. Example:
/apps/web/src/data/api_providers/openai.json

​
/apps/web/src/data/pricing

Tracks historical pricing data over time for trend analysis.
Files are timestamped and often aggregated by provider.

​
Documentation folders

Documentation is written in MDX and lives in /apps/docs/v1/.
Each folder maps directly to a group in the Mintlify configuration.
FolderPurpose
/apps/docs/v1/exploringExplains data concepts (models, benchmarks, providers).
/apps/docs/v1/contributingStep-by-step guides for contributors.
/apps/docs/v1/developersIntegration and API usage guides.
/apps/docs/v1/api-referenceDetailed reference for every endpoint.
/apps/docs/v1/communitySupport, FAQ, and community pages.

​
Application folders

PathDescription
/apps/webMain AI Stats website, built with Next.js 15 and TailwindCSS. Contains the data under src/data.
/apps/apiThe API Gateway - routes, caching, and model proxy logic.
/apps/docsThe Mintlify documentation project you’re reading now.
Each app runs independently but shares types and configuration through /packages.

​
SDK packages

The /packages directory contains SDKs for various programming languages to interact with the AI Stats API.
SDKDescription
sdk-tsTypeScript SDK for Node.js and web applications.
sdk-pyPython SDK.
sdk-goGo SDK.
sdk-csharpC# SDK.
sdk-phpPHP SDK.
sdk-rubyRuby SDK.
sdk-javaJava SDK.
sdk-rustRust SDK.
sdk-cppC++ SDK.
ai-sdk-ai-statsVercel AI SDK provider.

​
Automation and scripts

AI Stats uses automation to keep data up to date:
  • Importers - pull new data from sources like Hugging Face or provider APIs.
  • Watchers - run on a schedule (via GitHub Actions) to detect and commit updates.
  • Utilities - handle validation, schema checking, and JSON formatting.
You can explore these under /scripts/.

​
Where to start

If you’re contributing:
  • Data: Look in /apps/web/src/data/ for JSON files.
  • Documentation: Look in /apps/docs/v1/ for MDX files.
  • Code: Explore /apps/web or /apps/api.
  • Automation: Modify or inspect /scripts/.
  • SDKs: Check /packages/ for language-specific SDKs.

Editing a Model

Learn how to update or add model data.

Editing an Organisation

Understand how to modify organization data safely.
Last modified on February 11, 2026