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

# What is Theo?

> AI orchestration API — one call, every model, full pipeline.

Theo is an **AI orchestration engine** exposed as a developer API. Instead of calling individual model providers and building your own classification, routing, tool execution, and billing infrastructure, you call `POST /api/v1/completions` and Theo handles everything.

## What Theo Does

When you send a prompt to Theo, a multi-stage pipeline runs automatically:

1. **Intent Classification** — Determines what kind of task the prompt is (code, research, image, etc.)
2. **Skill Loading** — Loads domain expertise packages (installed + per-request) that inject system prompt extensions and tool definitions
3. **Model Routing** — Selects the optimal Theo engine for the task, with automatic failover
4. **Agent Loop** — Iteratively reasons, calls tools, and feeds results back until the task is complete
5. **Response** — Returns formatted content with engine info, tool execution log, artifacts, follow-up suggestions, and usage data

All of this happens in one API call.

## Key Capabilities

<CardGroup cols={2}>
  <Card title="Completions" icon="message-circle" href="/api-reference/completions/create">
    Multi-model AI completions with automatic intent classification and routing.
  </Card>

  <Card title="Skills" icon="puzzle" href="/skills/overview">
    Installable packages of domain knowledge and tools — like apps for AI.
  </Card>

  <Card title="E.V.I." icon="bot" href="/guides/build-an-evi">
    Embed Theo into your product with a custom persona. Your users see your brand, not Theo.
  </Card>

  <Card title="Code Generation" icon="code" href="/api-reference/code/generate">
    Theo Code engine with an extended output budget for long-form, best-in-class code.
  </Card>

  <Card title="Image Generation" icon="image" href="/api-reference/images/generate">
    Theo Create engine for photorealistic images with automatic fallback.
  </Card>

  <Card title="Research" icon="search" href="/api-reference/research/create">
    Deep web research with source synthesis and citations.
  </Card>

  <Card title="Documents" icon="file" href="/api-reference/documents/generate">
    Generate PDFs, DOCX, PPTX, XLSX, and CSV files.
  </Card>

  <Card title="Workflows" icon="workflow" href="/api-reference/workflows/create">
    Multi-step automations with schedules, event triggers, and conditional logic.
  </Card>
</CardGroup>

## Architecture at a Glance

```
Your App → @hitheo/sdk → Theo API (api.hitheo.ai)
                              │
                    ┌─────────┼─────────┐
                    ▼         ▼         ▼
              Classifier  Skill     Model
                          Loader    Router
                              │
                              ▼
                         Agent Loop
                    (think → act → observe)
                              │
                              ▼
                    Response + Billing + Audit
```

## Who It's For

* **Product developers** embedding AI into their apps via the SDK or REST API
* **Agencies and SaaS companies** building branded AI assistants (E.V.I.) for their customers
* **Skill authors** creating and monetizing domain-specific AI capabilities on the marketplace

## How to Get Started

<Steps>
  <Step title="Get an API key">
    Sign up at [api.hitheo.ai](https://api.hitheo.ai) and create a key from the dashboard. Keys are prefixed `theo_sk_`.
  </Step>

  <Step title="Make your first call">
    ```bash theme={null}
    curl -X POST https://www.hitheo.ai/api/v1/completions \
      -H "Authorization: Bearer theo_sk_..." \
      -H "Content-Type: application/json" \
      -d '{"prompt": "Explain quantum computing in one paragraph", "mode": "auto"}'
    ```
  </Step>

  <Step title="Install the SDK">
    ```bash theme={null}
    npm install @hitheo/sdk
    ```
  </Step>
</Steps>