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

# Media Analysis

> Analyze images, extract text from photos, transcribe audio and video, work with YouTube content, and find sounds — all from chat.

Pria can look at, listen to, and read your media. Upload an image and ask what's in it, drop in an audio or video file and get a transcript, pull a transcript from a YouTube link, or search a sound library — all without leaving the conversation.

## Image Analysis

Using a vision model, Pria can describe an image, identify objects, plants, or places, read the color palette, and extract text (OCR-style). Upload an image to your vault, or paste a public image URL, then ask.

```text Prompt wrap icon="comment" theme={null}
What's in this picture? <upload an image>
```

```text Prompt wrap icon="comment" theme={null}
Extract all the text from this photo of a sign.
```

<Tip>
  Try it for fun: upload a photo of a flower and ask which plant it is, or upload a landscape and ask Pria to guess the city or country.
</Tip>

## Transcribe Audio & Video

When you upload an audio or video file, Pria automatically extracts the speech as a text transcript and adds it to your vault. Large files are split and processed in parts. Once the transcript is ready, you can ask questions about what was said.

<Steps>
  <Step>Upload an audio or video file from the **\[+]** menu (file, URL, or drag-and-drop).</Step>
  <Step>Wait for the **Included** status — the audio has been transcribed and indexed.</Step>
  <Step>Ask questions about the content in plain language.</Step>
</Steps>

```text Prompt wrap icon="comment" theme={null}
What are the main points made in the recording I just uploaded?
```

## YouTube Transcripts & Playback

Give Pria a YouTube link and it can pull the transcript with timestamps, optionally save it to your vault, summarize it, or play the video inline at a specific moment.

```text Prompt wrap icon="comment" theme={null}
Transcribe and save the video at <youtube-link>
```

```text Prompt wrap icon="comment" theme={null}
Summarize the transcript of that video.
```

```text Prompt wrap icon="comment" theme={null}
Find where the video talks about "invisible axes" and play it.
```

<Note>
  The short share form of a link (the `youtu.be/…` URL) tends to work most reliably.
</Note>

## Play Audio & Video Files

When a file's URL is known — including ones you've uploaded — you can ask Pria to play it inline.

```text Prompt wrap icon="comment" theme={null}
Play my audio file.
```

```text Prompt wrap icon="comment" theme={null}
Play the video at <url> starting at 19 seconds.
```

## Find Sounds

Pria can search the FreeSound library for sound effects and clips.

```text Prompt wrap icon="comment" theme={null}
Find 3 sounds of tropical rain.
```

<img src="https://mintcdn.com/praxisai/RQ_2cJOeV_FmtYkN/images/user-guide/content-and-media/finding-sounds.webp?fit=max&auto=format&n=RQ_2cJOeV_FmtYkN&q=85&s=f43a103c64a084be69e70db4cdab5e57" alt="Finding Sounds" width="712" height="735" data-path="images/user-guide/content-and-media/finding-sounds.webp" />

## Tips & Limits

* Audio and video files are subject to upload size limits (25 MB per audio/video file); very large media is automatically split into parts.
* Transcription accuracy depends on audio quality — clear speech with minimal background noise transcribes best.
* Transcribing media and analyzing images consume credits, like any AI action.

## Related

* [Managing Files](/mdx/user-guide/files-documents/managing-files) — uploading media and the file action menu
* [Audio Notes](/mdx/user-guide/files-documents/audio-notes) — record voice memos that get transcribed into your vault
* [Generating Images](/mdx/user-guide/interface/image-generation) — create images from a description