> ## Documentation Index
> Fetch the complete documentation index at: https://docs.shannon.run/llms.txt
> Use this file to discover all available pages before exploring further.

# Components

> Shannon system components and their interactions

## System Components

<Note>
  This page is under construction as part of Phase 2 documentation expansion.
</Note>

Shannon is built as a distributed system with multiple specialized components:

## Core Services

### Gateway (Go)

* **Port**: 8080
* **Role**: REST API gateway, authentication, rate limiting
* **Technology**: Go, Chi router
* **Scaling**: Stateless, horizontal

### Orchestrator (Go)

* **Port**: 50052 (gRPC), 8081 (admin)
* **Role**: Task coordination, workflow management
* **Technology**: Go, Temporal, gRPC
* **Scaling**: Stateful via Temporal workflows

### Agent Core (Rust)

* **Port**: 50051 (gRPC)
* **Role**: Agent execution, tool invocation, WASI sandboxing
* **Technology**: Rust, Wasmtime, gRPC
* **Scaling**: Horizontal

### LLM Service (Python)

* **Port**: 8000
* **Role**: LLM provider gateway, prompt management
* **Technology**: Python, FastAPI, litellm
* **Scaling**: Horizontal

<Note>
  LLM Service notes:

  * GPT‑5 models are routed via the OpenAI Responses API, and the server prefers `output_text` when present. This avoids empty results when chat responses return structured content.
  * The `/agent/query` HTTP endpoint expects `allowed_tools`:
    * Omit/`null`: role presets may enable tools
    * `[]`: tools disabled
    * `["name", …]`: only listed tools available (built‑in, OpenAPI, MCP by registry name)
</Note>

### Desktop App (Tauri + Next.js)

* **Port**: — (native app, connects to Gateway :8080)
* **Role**: Real-time monitoring UI (tasks, runs, events)
* **Technology**: Next.js, Tauri, SSE/WebSocket
* **Scaling**: Client-side; backend services scale independently

## Data Stores

### PostgreSQL

* **Port**: 5432
* **Purpose**: Task metadata, events, sessions, users
* **Schema**: \~15 tables (task\_executions, event\_logs, sessions, etc.)

### Redis

* **Port**: 6379
* **Purpose**: Session cache, pub/sub, rate limiting
* **Persistence**: Optional (RDB/AOF)

### Qdrant

* **Port**: 6333
* **Purpose**: Vector embeddings, semantic memory
* **Collections**: Per-tenant isolation

### Temporal

* **Port**: 7233 (gRPC), 8088 (UI)
* **Purpose**: Durable workflow execution
* **Persistence**: PostgreSQL backend

## Component Interactions

<img src="https://mintcdn.com/ptmind-3aa3d4e1/2Bzddmlzr-QTR0Yc/en/architecture/assets/component-interactions.svg?fit=max&auto=format&n=2Bzddmlzr-QTR0Yc&q=85&s=191dd8d58ba25a6bb6b7651de5361b49" alt="Component Interactions" width="750" height="650" data-path="en/architecture/assets/component-interactions.svg" />

## Coming Soon

Detailed documentation for each component including:

* Architecture diagrams
* API interfaces
* Configuration options
* Scaling guidelines
* Troubleshooting tips

See [Architecture Overview](/en/architecture/overview) for high-level system design.