> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.vapi.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.vapi.ai/_mcp/server.

# What's New: Week of November 9, 2026

## New Models and Transcribers

1. **Reson8 transcriber**: Reson8 is now available as a transcriber, with support for multiple candidate languages, keyterms, custom models, and turn detection settings. You can use Vapi's key or your own Reson8 API key.
2. **Soniox stt-rt-v6 with assistant context**: Soniox's [`stt-rt-v6`](/providers/transcriber/soniox#supported-models) transcription model is now available. As the assistant speaks, Vapi streams its words to Soniox, so Soniox can filter out echo and understand the caller in the context of the conversation. The result is more accurate transcripts and fewer misheard turns.
3. **Claude Sonnet 4.6 on Bedrock**: Claude Sonnet 4.6 is now available through Amazon Bedrock's global cross-region profile (`global.anthropic.claude-sonnet-4-6`) when you use your own [AWS credentials](/providers/model/anthropic-bedrock).

## Features

4. **Background sounds in simulations**: You can now add a [background sound to a simulation personality](/observability/simulations-configure-ai-tester#background-sound) to test how your assistant handles a caller in a noisy place. Choose office, white noise, city, crowd, or farm, or use your own audio URL. You can also set the four new sounds, `white-noise`, `city`, `crowd`, and `farm`, with `backgroundSound` on an assistant.