AI Atlas
AI Atlas is a public web app for exploring real-world AI deployments across companies, industries, and countries.
The app combines an interactive globe, a searchable use case index, industry pages, an AI news digest, a data quality dashboard, and a blog/weekly report system backed by Supabase.
Live app: https://ai-atlas.app
What It Tracks
AI Atlas primarily tracks deployed or announced AI use cases. Each use case record is intended to describe:
- the organization using or building the system
- the industry or sector involved
- the country, city, or deployment location
- the source URL or reference material
- the use case description and related metadata
Data is collected and structured by an automated research pipeline, then published to the app for browsing and analysis.
The /news module adds a separate industry-signal layer: curated AI media and builder updates are summarized, tagged, and connected back to published AI Atlas use cases for editorial context.
Main Features
Interactive Globe
The home page shows companies and use cases on a 3D globe. Users can search, filter, select map markers, and open detail panels for individual organizations or deployments.
Use Case Index
The /use-cases page provides a searchable and filterable table of all tracked deployments. It supports keyword search, industry filters, country filters, sorting, pagination, column visibility, and detail modals.
AI News
The /news page curates AI industry updates from mainstream AI media and builder channels. News cards include source summaries, tags, source images or branded fallbacks, and an AI Atlas take that can reference published use cases. Use case references open the same detail modal used in the use case index, so readers can inspect related deployments without leaving the news page.
The news feed supports keyword search, source filters, tag filters, sorting, top and bottom pagination controls, selectable page size, and a floating Top shortcut. External news links open in a new browser tab/window.
News images are resolved through /api/news-image, which reads Open Graph/Twitter image metadata and serves images through a same-origin proxy. YouTube links use a stable video thumbnail fallback before proxying, which avoids browser-specific thumbnail failures.
Industry Explorer
The /industries pages group use cases by industry and summarize activity by company, country, recent deployments, and related weekly reports.
Blog And Weekly Reports
The /blog pages render analysis posts and generated weekly reports from Supabase. Weekly reports include highlights, trends, related case IDs, tags, and summary statistics.
Data Quality Dashboard
The /quality page provides a live data quality dashboard for AI Atlas records. It evaluates use cases and companies across completeness, validity, consistency, uniqueness, and traceability checks, then reports overall scores, table-level scores, issue severity, sample failures, status distributions, and metadata distributions.
Tech Stack
| Layer | Technology |
|---|
| App framework | Next.js App Router |
| Language | TypeScript |
| UI | React, Tailwind CSS, shadcn/ui, Radix UI |
| Globe | Three.js, react-globe.gl |
| Tables | TanStack Table |
| Database | Supabase Postgres |
| Deployment | Vercel |
Project Structure
app/ Next.js routes, layouts, pages, and API handlers
components/ Reusable UI and feature components
components/globe/ Interactive globe implementation
components/use-cases/ Use case table and detail modal components
components/news/ AI news feed, cards, image fallbacks, and AI Atlas take rendering
components/blog/ Blog article and list components
components/industries/ Industry summary and detail components
lib/ Data access, Supabase clients, types, utilities
hooks/ Shared React hooks
scripts/ Maintenance and content generation scripts
supabase/migrations/ Database migrations and policies
Core Routes
| Route | Purpose |
|---|
/ | Globe-based exploration experience |
/use-cases | Full use case index |
/use-cases/[id] | Shareable use case detail page |
/industries | Industry summary index |
/industries/[industry] | Industry detail page |
/news | AI news digest with source/tag filters and AI Atlas take links |
/blog | Blog and weekly report index |
/blog/[slug] | Blog post or weekly report detail |
/quality | Live data quality dashboard |
/api/news-image | Same-origin image metadata and thumbnail proxy for news cards |
/api/quality | JSON endpoint for dashboard quality checks |
/api/use-cases/[id] | JSON endpoint for a single normalized use case |
Data Model
The app primarily reads from these Supabase tables:
AI_Atlas_CompaniesAI_Atlas_Use_CasesAI_Atlas_NewsAI_Atlas_Blog_Posts
The data layer is implemented in:
lib/data.ts for companies, use cases, coordinates, and latest update timestampslib/data-news.ts for AI news reads from AI_Atlas_Newslib/data-blog.ts for blog and weekly report readslib/data-industries.ts for industry aggregationlib/blog-admin.ts for weekly report upserts
Server-side reads prefer SUPABASE_SERVICE_ROLE_KEY when available so the app can read data even when row-level security policies are incomplete. Public browser code only uses the anon key.
AI_Atlas_News stores curated news records with fields such as title, summary, url, source_name, published_at, tags, ai_atlas_take, and take_enriched_at. company_id is optional because many news items are industry signals rather than company-specific deployment records.
Prerequisites
- Node.js 18.17 or newer
- npm
- A Supabase project
Local Setup
Install dependencies:
npm install
Create .env.local in the repository root:
NEXT_PUBLIC_SUPABASE_URL=your_supabase_project_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key
SUPABASE_SERVICE_ROLE_KEY=your_service_role_key
SUPABASE_SERVICE_ROLE_KEY is only needed for server-side admin reads and maintenance scripts. Do not expose it in client code or commit it to git.
Run the development server:
npm run dev
Open:
http://localhost:3000
Environment Variables
| Variable | Required | Purpose |
|---|
NEXT_PUBLIC_SUPABASE_URL | Yes | Supabase project URL |
NEXT_PUBLIC_SUPABASE_ANON_KEY | Yes | Public anon key used by the app |
SUPABASE_SERVICE_ROLE_KEY | Recommended | Server-only key for privileged reads and maintenance scripts |
NEXT_DEV_LAN_ORIGINS | Optional | Comma-separated dev origins for testing from another device on the LAN |
BLOG_DISABLE_WEEKLY_STUB | Optional | Weekly report stubs are served in next dev only, never in production; set to true or 1 to disable them in dev too |
Available Scripts
npm run dev # Start the Next.js dev server
npm run build # Build the production app
npm run start # Start the production server
npm run lint # Run ESLint
AI News Ingestion
The news table is designed to be updated by an external agent pipeline such as OpenClaw. Each item should include an original, editorial-style summary rather than copied source text, plus tags and an optional ai_atlas_take generated by a separate agent. The app displays a disclosure on the news page:
Summaries are AI-assisted digests. Read the original source for full context.
When ai_atlas_take mentions published use cases by title, the news card attempts to link those references to the use case detail modal.
Data Quality Dashboard
Open the dashboard locally at:
http://localhost:3000/quality
The dashboard calls /api/quality, reads use case and company data from Supabase, and runs rule-based checks for required fields, source URLs, taxonomy validity, country/continent consistency, duplicate-like records, and traceability. It is intended as an operational review surface before publishing or expanding the dataset.
Database Setup
Database migrations live in supabase/migrations.
Typical Supabase workflow:
npx supabase link --project-ref your-project-ref
npx supabase db push
Use the Supabase SQL editor if you prefer applying individual migration files manually.
Security Notes
.env.local is ignored by git and should contain all local secrets.SUPABASE_SERVICE_ROLE_KEY must remain server-only./api/use-cases/[id] is a public JSON endpoint. Only expose fields that are safe for public readers./api/news-image fetches and proxies public source images for display; it should not be used for private or authenticated URLs./api/quality reports aggregate data quality findings and sample issue labels; avoid adding private fields to that response.- Avoid committing local database files, generated reports, private paths, or machine-specific configuration unless they are intentionally part of the public project.
Data Disclaimer
AI Atlas data is compiled from public sources such as company websites, announcements, public reports, industry media, builder channels, and related materials. Records and AI-assisted summaries are best-effort and may contain incomplete or outdated information. Verify important details against original sources before relying on them.
Company names, logos, and trademarks belong to their respective owners.
License
MIT
