[Stream] Audio Mode for Media Transformations (#23883)

* Media Transformations Audio Mode

* Update src/content/changelog/stream/2025-07-11-media-transformations-audio-mode.mdx

Co-authored-by: Taylor Smith <[email protected]>

* Update src/content/changelog/stream/2025-07-11-media-transformations-audio-mode.mdx

Co-authored-by: Taylor Smith <[email protected]>

* Update src/content/docs/stream/transform-videos/index.mdx

Co-authored-by: Taylor Smith <[email protected]>

* Update src/content/docs/stream/transform-videos/index.mdx

Co-authored-by: Taylor Smith <[email protected]>

* Roll forward changelog date

---------

Co-authored-by: Pakhi Sinha <[email protected]>
Co-authored-by: Pakhi <[email protected]>

Author: 3 people

src/content/changelog/stream/2025-07-22-media-transformations-audio-mode.mdx

@@ -0,0 +1,18 @@
+---
+title: Audio mode for Media Transforamtions
+description: >
+  Media Transformations now supports `audio` mode, which extracts audio from a video.
+date: 2025-07-22
+---
+
+The addition of this feature allows a user to extract audio from a source video, outputting
+an M4A file to use in downstream workflows like AI inference, content moderation, or transcription.
+
+For example,
+
+``` text title="Example URL"
+https://example.com/cdn-cgi/media/<OPTIONS>/<SOURCE-VIDEO>
+https://example.com/cdn-cgi/media/mode=audio,time=3s,duration=60s/<input video with diction>
+```
+
+For more information, learn about [Transforming Videos](/stream/transform-videos/).

src/content/docs/stream/transform-videos/index.mdx

@@ -54,12 +54,13 @@ Specifies the kind of output to generate.
 - `video`: Outputs an H.264/AAC optimized MP4 file.
 - `frame`: Outputs a still image.
 - `spritesheet`: Outputs a JPEG with multiple frames.
+- `audio`: Outputs an AAC encoded M4A file.
 
 ### `time`
 
 Specifies when to start extracting the output in the input file. Depends on `mode`:
 
-- When `mode` is `spritesheet` or `video`, specifies the timestamp where the output will start.
+- When `mode` is `spritesheet`, `video`, or `audio`, specifies the timestamp where the output will start.
 - When `mode` is `frame`, specifies the timestamp from which to extract the still image.
 - Formats as a time string, for example: 5s, 2m
 - Acceptable range: 0 – 10m
@@ -69,7 +70,7 @@ Specifies when to start extracting the output in the input file. Depends on `mod
 
 The duration of the output video or spritesheet. Depends on `mode`:
 
-- When `mode` is `video`, specifies the duration of the output.
+- When `mode` is `video` or `audio`, specifies the duration of the output.
 - When `mode` is `spritesheet`, specifies the time range from which to select frames.
 - Acceptable range: 1s - 60s (or 1m)
 - Default: input duration or 60 seconds, whichever is shorter
@@ -102,12 +103,18 @@ When `mode` is `video`, specifies whether or not to include the source audio in
 - `false`: Output will be silent.
 - Default: `true`
 
+When `mode` is `audio`, audio cannot be false.
+
 ### `format`
 
 If `mode` is `frame`, specifies the image output format.
 
 - Acceptable options: `jpg`, `png`
 
+If `mode` is `audio`, specifies the audio output format. 
+
+- Acceptable options: `m4a` (default)
+
 ## Source video requirements
 
 Input video must be less than 100MB.
@@ -128,6 +135,6 @@ Media Transformations will be free for all customers while in beta.
 After that, Media Transforamtions and Image Transformations will use the same subscriptions and usage metrics.
 
 - Generating a still frame (single image) from a video counts as 1 transformation.
-- Generating an optimized video counts as 1 transformation _per second of the output_ video.
+- Generating an optimized video or extracting audio counts as 1 transformation _per second of the output_ content.
 - Each unique transformation is only billed once per month.
 - All Media and Image Transformations cost $0.50 per 1,000 monthly unique transformation operations, with a free monthly allocation of 5,000.