# Aria

> Aria is an AI model gateway and managed agent runtime platform. It gives builders one OpenAI-compatible API for model calls, one dashboard for deploying Hermes agents, and one wallet for model and runtime usage.

## Primary URLs

- Homepage: https://entelic.io/
- Pricing and model catalog: https://entelic.io/pricing
- Model API docs: https://entelic.io/model-api
- Login / dashboard entry: https://entelic.io/login
- AI context: https://entelic.io/ai-context.md
- Full AI brief: https://entelic.io/llms-full.txt
- Structured product facts: https://entelic.io/product.json

## What Aria Does

- Provides an OpenAI-compatible Model API for direct model calls.
- Maintains a model catalog with route-aware prices, model availability, and billing snapshots.
- Deploys and manages Hermes agents through a hosted runtime control plane.
- Lets users choose platform-created model API keys when deploying agents or assembling AI workflows.
- Tracks usage by source, API key, runtime, and agent.
- Uses a unified wallet for direct model calls, runtime usage, and agent-driven model requests.
- Supports a skill ecosystem with Skills, Packs, and Assembly for reusable agent capabilities.
- Manages credentials through Connections instead of storing secrets in runtime configuration.

## Product Modules

- Model Gateway: API keys, OpenAI-compatible endpoints, model catalog, route groups, pricing, usage logs.
- Managed Agent Runtime: deploy, wake, monitor, and control Hermes agents.
- Skills: reusable capabilities that can be installed or assembled into agents.
- Packs: curated bundles of skills, prompts, and connection requirements for a use case.
- Assembly: an AI-assisted workflow for turning a goal into an agent configuration using user-owned model keys.
- Connections: secure credential storage and injection for channels, APIs, and tools.
- Billing Wallet: balance, holds, charges, refunds, and attribution for model and runtime consumption.

## Target Users

- Developers building AI products or internal tools.
- Founders who need hosted agents without managing runtime infrastructure.
- Teams that want transparent model usage, route-aware pricing, and runtime controls.
- Builders moving from prototypes to production agent operations.

## Positioning

Aria is not just a chatbot builder. It combines model access, runtime deployment, skills, credentials, billing, and usage observability in one control plane. It is closer to an agent infrastructure platform than a no-code chat UI.

## For AI Agents and Crawlers

The public web app is a React SPA. If a tool cannot execute JavaScript, use these static files:

- /llms.txt for a short index.
- /llms-full.txt for the detailed product brief.
- /ai-context.md for market positioning and messaging.
- /product.json for structured facts.
- /sitemap.xml for public routes.