# Orellius Docs

## Docs

- [Orellius Docs](https://docs.orellius.ai/introduction.md): Documentation for Orellius open-source projects.
- [Claude (Anthropic)](https://docs.orellius.ai/laminae/backends/claude.md): First-party EgoBackend for Claude via the Anthropic API.
- [Ollama (Local)](https://docs.orellius.ai/laminae/backends/ollama.md): Standalone HTTP client for Ollama. Used internally by Psyche, Persona, Cortex, and Shadow.
- [OpenAI / Compatible](https://docs.orellius.ai/laminae/backends/openai.md): EgoBackend for any OpenAI-compatible API: OpenAI, Groq, Together, DeepSeek, or local servers.
- [Architecture](https://docs.orellius.ai/laminae/getting-started/architecture.md): Crate structure, dependency graph, extension points, and data flow.
- [Installation](https://docs.orellius.ai/laminae/getting-started/installation.md): Add Laminae to your Rust project — full stack or individual layers.
- [Quick Start](https://docs.orellius.ai/laminae/getting-started/quickstart.md): Run the Psyche cognitive pipeline with a mock LLM in 5 minutes.
- [Introduction](https://docs.orellius.ai/laminae/introduction.md): The missing layer between raw LLMs and production AI.
- [Cortex](https://docs.orellius.ai/laminae/layers/cortex.md): Track how users edit AI output and convert corrections into reusable instructions — without fine-tuning.
- [Glassbox](https://docs.orellius.ai/laminae/layers/glassbox.md): Rust-enforced containment that no LLM can reason its way out of.
- [Ironclad](https://docs.orellius.ai/laminae/layers/ironclad.md): Three hard constraints enforced on all spawned sub-processes: command whitelist, platform-native sandbox, and resource watchdog.
- [Persona](https://docs.orellius.ai/laminae/layers/persona.md): Extract a writing personality from text samples and enforce it on LLM output.
- [Psyche](https://docs.orellius.ai/laminae/layers/psyche.md): A Freudian-inspired multi-agent architecture where three agents shape every response.
- [Shadow](https://docs.orellius.ai/laminae/layers/shadow.md): Automated security auditor that red-teams every AI response. Runs async — never blocks the conversation.
- [Full Stack Pipeline](https://docs.orellius.ai/laminae/recipes/full-stack.md): All six layers working together in a complete AI pipeline.
- [Safe Code Execution](https://docs.orellius.ai/laminae/recipes/safe-execution.md): Combine Glassbox and Ironclad for safe AI-driven code execution.
- [Voice-Matched Responses](https://docs.orellius.ai/laminae/recipes/voice-matching.md): Use Persona extraction + voice filter to make any LLM write like a specific person.
- [Benchmarks](https://docs.orellius.ai/laminae/reference/benchmarks.md): Performance numbers for every layer, measured with Criterion.rs on Apple M4 Max.
- [Changelog](https://docs.orellius.ai/laminae/reference/changelog.md): Release history for Laminae.
- [Configuration](https://docs.orellius.ai/laminae/reference/configuration.md): Configuration options for Shadow, Ironclad, Glassbox, Persona, and Ollama.
- [Extension Points](https://docs.orellius.ai/laminae/reference/extension-points.md): Every layer exposes a trait you can implement to customize behavior.
- [Python Bindings](https://docs.orellius.ai/laminae/reference/python-bindings.md): Use Glassbox, VoiceFilter, and Cortex from Python via PyO3.
- [WASM Support](https://docs.orellius.ai/laminae/reference/wasm.md): Run Glassbox, Persona voice filter, and Cortex in the browser via WebAssembly.
- [Architecture Overview](https://docs.orellius.ai/mcp-dome/architecture/overview.md): How MCPDome's interceptor chain processes every MCP message
- [Installation](https://docs.orellius.ai/mcp-dome/getting-started/installation.md): Install MCPDome from crates.io or build from source
- [Quick Start](https://docs.orellius.ai/mcp-dome/getting-started/quick-start.md): Get MCPDome running in under a minute
- [MCPDome](https://docs.orellius.ai/mcp-dome/introduction.md): Protective Dome for AI Agents — MCP security gateway proxy
- [Writing Policy Rules](https://docs.orellius.ai/mcp-dome/policy/writing-rules.md): Configure default-deny authorization with TOML policy rules

## OpenAPI Specs

- [openapi](https://docs.orellius.ai/api-reference/openapi.json)

## Optional

- [GitHub](https://github.com/orellius)
- [Blog](https://orellius.ai/blog)