> ## Documentation Index
> Fetch the complete documentation index at: https://comis-feature-matrix-channel.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Media

> Test media processing capabilities and configure media providers

The Media section provides two views: a **test playground** for verifying media processing capabilities, and a **configuration panel** for reviewing and managing media provider settings. Together, they let you confirm that vision, speech, documents, and other media features are working correctly before relying on them in production.

## Routes

* `/media` -- opens the Media Test playground
* `/media/config` -- opens the Media Configuration panel

## Backing RPC

* `media.providers` -- which providers are configured for each capability and whether the credentials work
* `media.transcribe`, `media.synthesize`, `media.analyzeImage`, `media.describeVideo`, `media.extractDocument`, `media.fetchLink` -- one RPC per capability
* `POST /media/upload` (HTTP, not RPC) -- multipart upload for audio/image/document files
* `config.patch(section: "media", ...)` -- save provider/model changes via the **Edit in Config Editor** links

Test calls are billed against your provider quota -- they hit the real API and produce real costs.

## What You See

The Media section is split into two views accessible from the sidebar under a single "Media" item.

### Media Test Playground

The test view has six tabs, one for each media processing capability:

| Tab          | What It Tests                                                                                                |
| ------------ | ------------------------------------------------------------------------------------------------------------ |
| **STT**      | Speech-to-Text -- upload an audio file (up to 25 MB) and get a transcription with provider info              |
| **TTS**      | Text-to-Speech -- enter text, synthesize speech, and play the audio in your browser                          |
| **Vision**   | Image analysis -- upload an image (up to 20 MB) and get a description or analysis from the vision provider   |
| **Document** | Document extraction -- upload a PDF, CSV, or other document (up to 50 MB) and get the extracted text content |
| **Video**    | Video analysis -- upload a video file (up to 50 MB) and get a description or analysis                        |
| **Link**     | Link enrichment -- enter a URL and get the extracted content, title, and metadata                            |

### How Testing Works

Each tab follows the same pattern:

1. Select or upload the input (file upload or text/URL input)
2. Click the test button
3. View the result, which includes the processing output and provider information

The view checks media provider availability on load and shows which providers are configured and ready. If a provider is not available, the corresponding tab will indicate that the capability is not configured.

### Provider Info

At the top of the test view, a summary shows which media providers are currently configured. This is fetched from the `media.providers` RPC endpoint and shows the active provider for each capability (e.g., OpenAI for STT, ElevenLabs for TTS).

### Media Configuration

The configuration panel (at `/media/config`) shows provider status and current settings for all five media subsystems:

| Subsystem               | What It Configures                                                             |
| ----------------------- | ------------------------------------------------------------------------------ |
| **STT**                 | Speech-to-Text provider (OpenAI, Groq, Deepgram), model selection, language    |
| **TTS**                 | Text-to-Speech provider (OpenAI, ElevenLabs, Edge TTS), voice selection, speed |
| **Vision**              | Vision analysis provider, model, max tokens for descriptions                   |
| **Document Extraction** | PDF and document processing settings                                           |
| **Link Understanding**  | Web content extraction and readability settings                                |

Each subsystem is displayed as a card with:

* **Status indicator** -- whether the provider is configured and operational
* **Current settings** -- the active provider, model, and key configuration values
* **Edit link** -- a "Edit in Config Editor" link that navigates to the relevant config section for detailed editing

## Common Tasks

<Steps>
  <Step title="Verify STT is working">
    Go to `/media`, select the STT tab, upload an audio file, and click Test. The transcription should appear with the provider name.
  </Step>

  <Step title="Test TTS synthesis">
    Go to `/media`, select the TTS tab, enter some text, and click Synthesize. An audio player should appear letting you play the generated speech.
  </Step>

  <Step title="Check media provider configuration">
    Go to `/media/config` to see all five media subsystems with their status and settings. Use the "Edit in Config Editor" links to change providers.
  </Step>

  <Step title="Test vision analysis">
    Go to `/media`, select the Vision tab, upload an image, and click Analyze. The analysis result shows what the vision model detected in the image.
  </Step>
</Steps>

## Related Pages

<CardGroup cols={2}>
  <Card title="Media & Voice Overview" icon="microphone" href="/media/index">
    Full documentation for all media capabilities.
  </Card>

  <Card title="Vision" icon="eye" href="/media/vision">
    Detailed vision processing documentation.
  </Card>

  <Card title="Voice" icon="volume-high" href="/media/voice">
    Speech-to-text and text-to-speech documentation.
  </Card>

  <Card title="Config Editor" icon="gear" href="/web-dashboard/config-editor">
    Edit media provider configuration in the YAML editor.
  </Card>
</CardGroup>