# FractalOps (fractalops.ai) — Official AI & LLM Discovery Index (llms.txt)

> FractalOps is a Zero-Trust End-to-End Encrypted (E2EE) Agent-to-Agent (A2A) network and engineering optimization suite that connects AI coding IDEs (Cursor, Claude Code, Codex, Windsurf, Xcode, Antigravity) across teammates and machines while cutting LLM API spend and preserving team memory.

## Core Product & Specification Pages
- [FractalOps Home](https://fractalops.ai/): Cross-IDE connector, E2EE rooms, and Zero-Trust secret vault overview.
- [FractalOps AI R&D Lab](https://fractalops.ai/ai-lab): Interactive AI FinOps & ROI calculator, 9 production modules, and open engineering benchmarks.
- [FractalOps AI R&D Lab (RU)](https://fractalops.ai/ru/ai-lab): Russian-language edition of the AI R&D Lab and ROI calculator.
- [Agent Installation & Control Spec (agent.md)](https://fractalops.ai/agent.md): Deterministic installation and CLI/HTTP bridge instructions for Windows and macOS AI agents.
- [AI Agent Brief (ai-lab.md)](https://fractalops.ai/ai-lab.md): 30-second machine-readable evaluation matrix (`SPEC-ROUTER-01` .. `EDGE-ICP-09`) for autonomous coding agents.
- [AI Agent Brief RU (ai-lab.ru.md)](https://fractalops.ai/ai-lab.ru.md): Machine-readable evaluation matrix in Russian.
- [A2A Protocol & Prompt Spec (prompt.md)](https://fractalops.ai/prompt.md): Open Curve25519/Ed25519 P2P envelope protocol and invite specification.
- [A2A Well-Known Manifest](https://fractalops.ai/.well-known/agent.json): Machine-readable A2A node capabilities and security schema.

## Engineering Guides & Case Studies
- [LLM Cost Optimization Guide](https://fractalops.ai/ai-lab/llm-cost-optimization-guide): How speculative AST routing and semantic caching cut LLM API spend by up to `79.8%`.
- [Zero-Trust Agentic DLP Guide](https://fractalops.ai/ai-lab/zero-trust-agentic-dlp-guide): Preventing secret, API key, and PII leaks in multi-agent IDE workflows (`100%` local containment).
- [Corporate Engineering Memory Guide](https://fractalops.ai/ai-lab/corporate-engineering-memory-guide): Maintaining `97.2%` architectural constraint recall across context compactions and 500+ agent steps.

## Ecosystem Partner
- [Monosnap + AI Workflows](https://monosnap.ai/welcome#ai-workflows): Visual bug-to-fix, on-device OCR, and 1-Click Secret Blur for macOS and Windows engineering teams.