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

# CDN & Static Assets

> Access static JSON, audio, entities, and embeddings from cdn.urantia.dev — the Cloudflare R2-backed CDN powering the Urantia Papers API.

All static data that powers the API is available directly from a public Cloudflare R2 bucket at `cdn.urantia.dev`. This includes the audio files. You can use these files for offline apps, custom pipelines, or self-hosted setups.

<Warning>
  Audio moved from the old `audio.urantia.dev` subdomain to `cdn.urantia.dev/audio/...`. The old host is frozen and serves only part of the library. Update any saved URLs to the patterns below.
</Warning>

## What's available

| Asset | Format | Count | Description |
| - | - | - | - |
| Papers | JSON | 197 | Full paper text with paragraphs, sections, and metadata |
| Paper audio | MP3 | 197 | Full narration of each paper |
| Paragraph audio | MP3 | 16,000+ | Individual paragraph narration across 7 voices |
| Paper videos | MP4 | 197 | Full paper narration with synced text overlay (1080p) |
| Video thumbnails | PNG | 197 | Paper title cards for video thumbnails |
| Entities | JSON | 3,000+ | Typed entities (beings, places, concepts) with descriptions |
| Entity links | JSON | 50,000+ | Paragraph-to-entity citation mappings |
| Embeddings | JSON | 14,500+ | OpenAI `text-embedding-3-small` vectors (1536-dimensional) |
| Audio manifest | JSON | 1 | Complete inventory mapping paragraph IDs to audio URLs |

## URL patterns

### Audio files

Paragraph-level audio follows this naming pattern:

```
https://cdn.urantia.dev/audio/eng/paragraphs/{voice}/{model}-{voice}-{globalId}.mp3
```

**Examples:**

```bash theme={null}
# tts-1-hd model, nova voice, paragraph 0:0.0.1 (Foreword, first paragraph)
https://cdn.urantia.dev/audio/eng/paragraphs/nova/tts-1-hd-nova-0:0.0.1.mp3

# gpt-4o-mini-tts model, cedar voice, same paragraph
https://cdn.urantia.dev/audio/eng/paragraphs/cedar/gpt-4o-mini-tts-cedar-0:0.0.1.mp3

# Whole paper audio (Paper 1). Use 0 for the Foreword
https://cdn.urantia.dev/audio/eng/papers/1.mp3

# Part-level audio
https://cdn.urantia.dev/audio/eng/parts/tts-1-hd-nova-Part1.mp3
```

Only `nova` covers every paragraph. For any other voice, read the `audio` object on the paragraph instead of building a URL, because a missing voice returns a 404.

**Available voices:**

| Model | Voices | Coverage |
| - | - | - |
| `tts-1-hd` | `nova` | Full |
| `gpt-4o-mini-tts` | `cedar` | Partial |
| `tts-1-hd` | `onyx` | Partial |
| `tts-1-hd` | `echo`, `alloy`, `fable`, `shimmer` | Samples only |
| `tts-1` | `alloy` | Samples only |

### Video files

Paper-level videos with narrated audio and synced text overlay:

```bash theme={null}
# Paper 1 video (H.264 MP4, 1080p, 30fps)
https://video.urantiahub.com/tts-1-hd-nova-1.mp4

# Paper 1 thumbnail (PNG, 1920x1080)
https://video.urantiahub.com/thumbnail-1.png
```

All 197 papers have videos with the `nova` voice. Videos include animated background, paragraph text fading in/out, section title cards, and intro/outro branding.

### JSON files

Paper JSON files are numbered `000.json` through `196.json`:

```bash theme={null}
# Foreword
https://cdn.urantia.dev/json/eng/000.json

# Paper 1 — The Universal Father
https://cdn.urantia.dev/json/eng/001.json

# Paper index (all papers, sections, paragraphs)
https://cdn.urantia.dev/json/eng/index.json

# Part metadata
https://cdn.urantia.dev/json/eng/1-part.json
```

### Entities

```bash theme={null}
# All entities with descriptions, aliases, and types
https://cdn.urantia.dev/entities/seed-entities.json

# Paragraph-to-entity citation mappings
https://cdn.urantia.dev/entities/paragraph-entities.json
```

### Embeddings

```bash theme={null}
# All paragraph embeddings (OpenAI text-embedding-3-small, 1536 dimensions)
https://cdn.urantia.dev/embeddings/embeddings.json
```

<Warning>
  The embeddings file is \~455 MB. Only download it if you need vector search capabilities for a self-hosted setup.
</Warning>

### Audio manifest

```bash theme={null}
# Complete mapping of paragraph IDs to audio CDN URLs
https://cdn.urantia.dev/manifests/audio-manifest.json
```

## Paper JSON structure

Each paper file contains the full text with structured metadata:

```json theme={null}
{
  "paperId": "1",
  "paperTitle": "The Universal Father",
  "partId": "1",
  "sections": [
    {
      "sectionId": "0",
      "sectionTitle": "",
      "paragraphs": [
        {
          "globalId": "1:1.0.1",
          "standardReferenceId": "1:0.1",
          "paperSectionParagraphId": "1.0.1",
          "text": "THE Universal Father is the God of all creation...",
          "htmlText": "<p>THE Universal Father is the God of all creation...</p>"
        }
      ]
    }
  ]
}
```

## Use cases

* **Offline apps** — Download the JSON files and audio for offline reading and listening
* **Custom search** — Use the embeddings to build your own vector search
* **Self-hosted API** — Seed your own database from the source JSON and entities
* **Data analysis** — Analyze entity relationships, citation patterns, or text embeddings
* **Alternative TTS** — Use the text data to generate audio with other TTS providers

<Tip>
  For most use cases, the [API](/api-reference/introduction) is easier than working with raw files. Use the CDN when you need bulk access or offline capabilities.
</Tip>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.