Public web app for browsing ESMA MiCA registers in one searchable interface.

Live app: www.micaregister.com
Backend API docs (recommended custom domain): https://api.micaregister.com/docs

• Aggregates all 5 ESMA MiCA registers in one UI:
  • CASP (Crypto-Asset Service Providers)
  • OTHER (White papers for other crypto-assets)
  • ART (Asset-Referenced Token Issuers)
  • EMT (E-Money Token Issuers)
  • NCASP (Non-compliant entities)
• Fast filtering by name, country, LEI, services, dates, and register-specific fields
• Shareable entity deep links (for example /casp/{entity_id}) that open details directly
• CASP list and detail views group duplicate LEI rows into one company card with multiple authorisation records
• Home Member State filtering falls back to LEI country when ESMA country is missing
• Backend API for search, filtering, and admin imports
• Data update scripts for download, validation, cleaning, and import

• Backend: FastAPI, SQLAlchemy, Alembic-style migrations
• Frontend: React + Vite
• Database: PostgreSQL (production) or SQLite (local)
• Deploy: Railway (backend) + Vercel (frontend)

git clone https://github.com/Kymylyy/mica-register.git
cd mica-register

python3 -m venv .venv
source .venv/bin/activate
pip install -r backend/requirements.txt

Optional database config:

export DATABASE_URL="postgresql://user:password@localhost:5432/mica_register"

If DATABASE_URL is not set, local SQLite is used.

python3 backend/import_all_registers.py

cd backend
uvicorn app.main:app --reload

Backend URL: http://localhost:8000

cd frontend
npm install
npm run dev

Frontend URL: http://localhost:5173

Backend:

• DATABASE_URL (required in production)
• CORS_ORIGINS (comma-separated allowed origins)
• ADMIN_API_TOKEN (or ADMIN_TOKEN) for /api/admin/* and authenticated entity write endpoints
• DEEPSEEK_API_KEY (optional, only for LLM remediation flows)

Frontend:

• VITE_API_URL (e.g. https://api.micaregister.com in production)

See .env.example and frontend/.env.example for local templates.

• Connect repository to Railway
• Ensure PostgreSQL service is attached
• Set required env vars (DATABASE_URL, CORS_ORIGINS, ADMIN_API_TOKEN)
• Schema bootstrap in production happens via backend/start.sh, which runs the repo migrations before starting the API.
• Add custom domain (recommended): api.micaregister.com
• Optional cron: run python scripts/run_railway_cron_update.py

• Import frontend project
• Set VITE_API_URL to your API domain (recommended: https://api.micaregister.com)
• Build command: npm run build
• Output directory: dist
• Build output includes static prerendered pages for /, /casp, /other, /art, /emt, and /ncasp, plus entity detail pages under stable slugs (for example /casp/euwax-ag) based on the latest API snapshot available during build. Legacy numeric-id URLs (/casp/{id}) get redirect stubs pointing at the slug URL; unknown paths are served by a noindex SPA fallback (spa-fallback.html).
• sitemap.xml and robots.txt are generated during prerender from the same route set (including entity detail URLs when available at build time).
• micaregister.com is redirected permanently (301) to www.micaregister.com.
• Detail prerender validation: by default it fails on CI/Vercel (PRERENDER_DETAIL_FAILURE_MODE=error) and warns locally (warn) when detail pages cannot be generated or are below PRERENDER_MIN_DETAIL_PAGES (default 1).

Per-register feeds (public):

• GET /api/feeds (index with feed links + docs + OpenAPI URL on current host)
• GET /api/feeds/{register}.json
• GET /api/feeds/{register}.csv
• API crawler endpoints on the API host:
  • GET /robots.txt
  • GET /sitemap.xml

Where {register} is one of: casp, other, art, emt, ncasp.

CASP grouped UI behavior:

• The web UI groups CASP entries by LEI for list and detail views.
• Public CASP feed endpoints remain raw-row exports of the underlying ESMA-style data.

Operational behavior:

• Feed responses are cached in API process memory for 300 seconds.
• Feed responses include ETag and Cache-Control headers for conditional requests and client/proxy caching.
• Feed endpoints are rate limited to 30 requests per 60 seconds per client IP.

List endpoints support single-column global sorting before pagination:

• GET /api/entities
• GET /api/casp/companies

Query params:

• sort_by=<field>
• sort_dir=asc|desc

Behavior:

• Sorting is applied to the full filtered result set before skip and limit.
• When sort_by is omitted, each endpoint keeps its existing default ordering.
• Date fields always place null values at the end.
• Unsupported sort fields return 400.

Supported sort_by values depend on the register and match the sortable columns exposed in the UI for that register.

• GET /llms.txt is served by the frontend domain (https://www.micaregister.com/llms.txt).
• It documents canonical host, crawl entry points, entity URL pattern, API docs, data source notes, and technical contact.

Frontend:

• cd frontend && npm run lint
• cd frontend && npm run typecheck
• cd frontend && npm run test -- --run
• cd frontend && npm run build

Backend:

• python3 -m pytest

CI runs the same checks on every pull request and push.

Primary source: ESMA MiCA register publications.

Before redistributing data snapshots from data/, confirm legal/licensing requirements for your use case.

See CONTRIBUTING.md.

See SECURITY.md for vulnerability reporting and secret handling expectations.

MIT. See LICENSE.