|
| 1 | +## New Features |
| 2 | + |
| 3 | +### [Inworld voice integration](https://docs.vapi.ai/changelog/2025/6/24) |
| 4 | + |
| 5 | +- Added support for Inworld as a TTS provider. |
| 6 | + |
| 7 | +- Choose from a broad portfolio of Inworld voices (English, Mandarin, Dutch, French, German, Italian, Japanese, Korean, Polish, Portuguese, Spanish). |
| 8 | + |
| 9 | +- Bring your own Inworld API key, configure caching and fallback behavior via API. |
| 10 | + |
| 11 | +## Improvements |
| 12 | + |
| 13 | +- Added support for [LMNT voice language selection](https://docs.vapi.ai/changelog/2025/6/25), allowing you to specify the language for LMNT TTS voices. |
| 14 | + |
| 15 | +- [Enhanced campaign metrics with new counters](https://docs.vapi.ai/changelog/2025/6/25) for queued, scheduled, in-progress, and voicemail-ended calls. |
| 16 | + |
| 17 | +- Added the ability to include [custom metadata for MCP tools](https://docs.vapi.ai/changelog/2025/6/25), enabling more advanced tool integrations. |
| 18 | + |
| 19 | +- Expanded [customer metadata creation options](https://docs.vapi.ai/changelog/2025/6/26) with support for email and external customer IDs. |
| 20 | + |
| 21 | +- Improved assistant and workflow configuration with [more flexible credential and compliance plan options](https://docs.vapi.ai/changelog/2025/6/20). |
| 22 | + |
| 23 | +- Enhanced [chat and web chat APIs](https://docs.vapi.ai/changelog/2025/6/26) with new session and assistant override capabilities. |
| 24 | + |
| 25 | +## Bug Fixes |
| 26 | + |
| 27 | +- Fixed inconsistent application of [assistant variable substitution](https://docs.vapi.ai/changelog/2025/6/26) in web chat. |
| 28 | + |
| 29 | +- Resolved validation issue for non-E.164 numbers with [customer number formatting checks for non-E164 numbers](https://docs.vapi.ai/changelog/2025/6/26) during batch calls. |
| 30 | + |
| 31 | +- Addressed edge case where [fallback plans for voices could fail to trigger on provider errors](https://docs.vapi.ai/changelog/2025/6/25). |
| 32 | + |
| 33 | +## Behind The Scenes |
| 34 | + |
| 35 | +- Removed the [deprecated Deepgram code switching](https://docs.vapi.ai/changelog/2025/6/24) transcriber option. |
| 36 | + |
| 37 | +- Added API endpoints for [workflow generation](https://docs.vapi.ai/changelog/2025/6/25). |
| 38 | + |
| 39 | +- Introduced new workflow and assistant schemas for Google and custom LLM models via [new workflow and assistant schema definitions for Google and custom LLM models](https://docs.vapi.ai/changelog/2025/6/25). |
| 40 | + |
| 41 | +- Enabled [more granular assistant, workflow, and tool configuration](https://docs.vapi.ai/changelog/2025/6/20). |
| 42 | + |
| 43 | +- Enhanced [MCP tool protocol support](https://docs.vapi.ai/changelog/2025/6/25) with metadata options for streamable HTTP and SSE. |
| 44 | + |
| 45 | +- Expanded authentication with [support for HMAC authentication plans](https://docs.vapi.ai/changelog/2025/6/24). |
| 46 | + |
| 47 | +- Refactored [campaign and customer](https://docs.vapi.ai/changelog/2025/6/26) objects for improved scalability and clarity. |
| 48 | + |
| 49 | +- Expanded [observability and compliance configuration](https://docs.vapi.ai/changelog/2025/6/25) for organizations, assistants, and workflows. |
0 commit comments