# AvaScry — Magic: The Gathering Visual & Semantic Knowledge Corpus

> AvaScry is an AI-native semantic knowledge corpus and visual engine for Magic: The Gathering. Powered by 4096-dimensional Qwen 8B vector embeddings across 38,000+ Oracle cards, sub-millisecond card rendering, comprehensive rules layers, errata history, and format legalities.

## Corpus Metadata & Versioning
- **Corpus-Version:** 1.0
- **Markdown-Schema:** 1.0
- **Similarity-Model:** Qwen-8B-4096d
- **Update-Frequency:** Daily
- **Canonical-Host:** https://avascry.com

## Machine Ingestion Guarantees
- All `/printing/*.md` endpoints begin with structured YAML frontmatter.
- All `/similar/*.md` endpoints provide 4096-dimensional nearest neighbors in ranked order.
- All rules and errata align with the official Magic Comprehensive Rules (CR 117, 613, 614, 704).
- Zero hallucination: Exact Oracle texts, Mana Values, Power/Toughness, and Legalities.

## AI Agent Integration & Markdown Endpoints

AI agents, LLMs, and crawlers can retrieve token-efficient, dense Markdown payloads across all surfaces:

- **Card Printings (.md):** `https://avascry.com/printing/<card-slug>-<set>.md`
- **4096-Dim Neural Synergies (.md):** `https://avascry.com/similar/<card-slug>.md`
- **Artist Portfolios (.md):** `https://avascry.com/artist/<artist-slug>.md`
- **Sets Directory (.md):** `https://avascry.com/sets.md`
- **Comprehensive Rules & Errata (.md):** `https://avascry.com/rules.md`
- **Format Taxonomy & Banlists (.md):** `https://avascry.com/legalities.md`
- **Markdown Sitemap Index:** `https://avascry.com/sitemap.md`
- **HTML Sitemap Directory:** `https://avascry.com/sitemap.html`
- **XML Sitemap Index:** `https://avascry.com/sitemap.xml`
- **AI Content Manifest:** `https://avascry.com/.well-known/ai-content`

## Retrieval Examples

### Content Negotiation Header (Preferred)
```bash
curl -H "Accept: text/markdown" https://avascry.com/printing/combat-research-dmu
```

### Direct Markdown URL
```bash
curl https://avascry.com/printing/combat-research-dmu.md
curl https://avascry.com/similar/combat-research.md
curl https://avascry.com/rules.md
```

### Python Retrieval
```python
import httpx
client = httpx.Client()
resp = client.get("https://avascry.com/printing/sol-ring-lea.md")
print(resp.text)
```

## Preferred Crawl Sequence for Autonomous Agents
1. `https://avascry.com/llms.txt` (This manifest)
2. `https://avascry.com/rules.md` (Rules mechanics, continuous layers & priority)
3. `https://avascry.com/legalities.md` (Format constraints & banlists)
4. `https://avascry.com/sitemap.html` (Complete category crawl web)
5. `https://avascry.com/similar/<slug>.md` (4096-dim synergy graph)
6. `https://avascry.com/printing/<slug>.md` (Canonical card documents)

## Crawling & Usage Policy
- AI Search, Ingestion & Training: Allowed (GPTBot, OAI-SearchBot, ClaudeBot, Anthropic-AI, PerplexityBot, Applebot, Amazonbot, Googlebot, Bingbot)
- Permitted Routes: `/`, `/printing/*`, `/similar/*`, `/artist/*`, `/set/*`, `/sets`, `/rules.md`, `/legalities.md`, `/commander`
- Restricted Routes: `/api/*`, `/auth/*`, `/dashboard`