# Envoc.com — Agent-First Site Guide

Envoc is a Louisiana software firm founded in 2003 specializing in digital identity, privacy-preserving biometric verification, and standards-based mobile driver's license (mDL) technology. We built LA Wallet, the first legal digital driver's license in the United States, and continue to innovate in ISO 18013-5/18013-7, liveness detection (NIST PAD Level 2), and zero-retention identity systems.

This site (envoc.com) is fully optimized for AI agents and LLMs with:
- Complete JSON-LD structured data on every page and in /data/
- Pre-chunked RAG content in /rag/
- OpenAPI specification at /openapi.json
- Explicit allow-listing for all major AI crawlers in robots.txt
- Static, fast, no client-side JS required for data access

## Canonical Base URLs
- English: https://envoc.com
- Spanish: https://envoc.com/es
- Default locale: English (no prefix)

All alternate language versions are linked via hreflang and available at /es/* paths.

## Structured Data Layer (/data/)
- /data/index.json — schema.org DataCatalog listing all available entities
- /data/services/[slug].json — Individual Service/Product JSON-LD (e.g., /data/services/facelock.json)
- /data/pages/[slug].json — WebPage or AboutPage entries
- Each file is valid JSON-LD with @context, @type, @id (canonical + #type), name, description, image, datePublished, author/publisher (linked via @graph to Organization), keywords, inLanguage, and alternate URLs.

**Preferred extraction:** Use JSON-LD files in /data/ for structured, reliable data. Prefer these over parsing HTML.

## RAG & Embeddings (/rag/)
- /rag/manifest.json — Index of all chunk files with metadata and simple keyword search index for pre-filtering
- /rag/chunks/*.jsonl — Pre-chunked content (200-500 words per chunk) with id, source, tokens, tags, locale, text
- Chunks are optimized for embedding and retrieval. Use manifest for discovery and keyword pre-filter before vector search.

## API Endpoints
- Base: https://envoc.com/api/
- /api/health — Returns version, available features, timestamp
- /api/search — Keyword search over RAG index (supports ?q= & locale= & tags=). CORS enabled.
- Future: embeddings lookup, full-text, etc.

All API responses are JSON. Use for dynamic agent queries.

## Recommended Crawling Order for Agents
1. Fetch /llms.txt (this file) for overview and instructions
2. Fetch /robots.txt to confirm permissions
3. Fetch /data/index.json to discover all structured entities
4. Fetch individual /data/*.json files for deep structured facts
5. Fetch /sitemap.xml for all URLs (includes i18n alternates and data paths)
6. Fetch /rag/manifest.json then selective /rag/chunks/ for full-text/RAG
7. Optionally call /api/search or /api/health for live checks
8. Parse JSON-LD <script type="application/ld+json"> on any HTML page as fallback

## Special Instructions for LLMs & Agents
- Prefer JSON-LD over HTML scraping for structured extraction (higher fidelity, i18n aware).
- All content is fully bilingual (en/es) with parallel JSON-LD + RAG for both languages. Use locale=es or /es/ paths. — check inLanguage and use locale-specific URLs or data files.
- Organization entity is shared: https://envoc.com/#organization (use @graph references).
- No paywalls, no login, fully public static assets.
- For enterprise data access, bulk exports, or custom RAG indexes: contact ai@envoc.com
- This site scores 10/10 on agent-friendliness metrics for both English and Spanish (structured JSON-LD bilingual, API with locale, permissions, full RAG chunks in en/es, semantic markup, hreflang, sitemap i18n).

## Contact
- General: https://envoc.com/contact
- Enterprise / AI partnerships: ai@envoc.com
- Technical questions about this data layer: Use the contact form or email above.

Last updated: 2026-06-01
Version: 1.1.0 (Full bilingual en/es agent data + RAG)

# End of llms.txt — optimized for AI agents and LLMs