feat: streaming keyterms prompt (#398)

* update streaming info

* minor verbiage tweak

* minor updates on format_turns behaviour

Author: m-ods

fern/pages/02-speech-to-text/universal-streaming/universal-streaming-keyterms.mdx

@@ -3,9 +3,8 @@
 The keyterm prompting feature helps improve recognition accuracy for specific words and phrases that are important to your use case.
 
 <Warning>
-Keyterms Prompting is currently in beta and available to all users free of charge. Pricing is still being finalized and may apply in the future.
 
-As we continue to develop this feature, functionality may evolve. For the latest updates and code examples, please refer to this page.
+Keyterms Prompting costs an additional $0.04/hour.
 
 </Warning>
 
@@ -562,13 +561,25 @@ if __name__ == "__main__":
 To utilize keyterm prompting, you need to include your desired keyterms as query parameters in the WebSocket URL.
 
 - You can include a maximum of 100 keyterms per session.
-- Each individual keyterm string must be between 5 and 50 characters in length.
-- The `format_turns` parameter must be set to `True` for keyterm prompting to be applied.
+- Each individual keyterm string must be 50 characters or less in length.
+
+## How it works
+
+Streaming Keyterm Prompting has two components to improve accuracy for your terms.
+
+### Word-level boosting
+
+The streaming model itself is biased during inference to be more accurate at identifying words from your keyterm list. This happens in real-time as words are emitted during the streaming process, providing immediate improvements to recognition accuracy. This component is enabled by default.
+
+### Turn-level boosting
+
+After each turn is completed, an additional metaphone-based boosting pass analyzes the full transcript using your keyterm list. This post-processing step, similar to formatting, provides a second layer of accuracy improvement by examining the complete context of the turn. To enable this component, set `format_turns` to `True`.
+
+Both stages work together to maximize recognition accuracy for your keyterms throughout the streaming process.
 
 ## Important notes
 
-- Only final formatted transcripts receive keyterm prompting.
-- Keyterm phrases outside the 5-50 character range are ignored.
+- Keyterm prompts longer than 50 characters are ignored.
 - Requests containing more than 100 keyterms will result in an error.
 
 ## Best practices