# v0.6.1 Release Notes

> v0.6.0 -> v0.6.1

## Summary

Fixes speech model discovery against task filters and adds optional TTS output formats while keeping the existing default.

<details>
<summary>翻譯</summary>

修正語音模型依任務篩選的發現邏輯，並在維持既有預設格式下支援可選的語音輸出格式。

</details>

## Changes

### FIX
- Respect Cloudflare STT/TTS model filters by matching Automatic Speech Recognition and Text-to-Speech tasks instead of always requiring Text Generation (@pardnchiu) [c620ad7]

<details>
<summary>翻譯</summary>

- 修正 Cloudflare 的 STT／TTS 模型篩選，改依 Automatic Speech Recognition 與 Text-to-Speech 任務比對，不再一律要求 Text Generation

</details>

### UPDATE
- Add optional `TTSOptions.Format` for OpenAI speech output (`wav` default; `mp3`, `opus`, `aac`, `flac`, `pcm`) with matching MIME types (@pardnchiu) [c620ad7]

<details>
<summary>翻譯</summary>

- 新增可選的 `TTSOptions.Format` 供 OpenAI 語音輸出（預設 `wav`；可選 `mp3`、`opus`、`aac`、`flac`、`pcm`），並對應正確 MIME

</details>

### DOC
- Rewrite the documentation to reflect the current router architecture and package structure (@pardnchiu) [2f08cf5]

<details>
<summary>翻譯</summary>

- 重寫文件以反映目前的路由器架構與套件結構

</details>

## Scope
- `core/audio.go` — UPDATE
- `core/openai/` — UPDATE (`audio.go`)
- `doc/` — DOC (`README.zh.md`, `architecture.md`, `architecture.zh.md`, `doc.md`, `doc.zh.md`)

***

Generated by [SKILL](https://github.com/pardnchiu/skill-version-generate)