#5542 AttributeError: 'NoneType' object has no attribute 'keys'

The error occurs when switching the TTS channel – the rolelist variable is None, meaning the software failed to load the list of available voices for that channel.

Root cause:
Each TTS channel stores its voice list in a JSON file inside videotrans/voicejson/. If that file is missing, corrupted, or fails to download (e.g., due to network issues when fetching online voices like Edge-TTS), the software passes None instead of a dictionary.

Solutions (in order of likelihood):

1. Check voice JSON files

   • Go to videotrans/voicejson/ in your installation directory.
   • Look for a JSON file matching the TTS channel you selected (e.g., edge-tts.json, qwen-tts.json, etc.).
   • If missing or empty, try re-downloading a fresh copy from the GitHub repository.
   • If present but corrupt, delete it and restart the software – it will be regenerated.

2. Ensure network connectivity

   • For online TTS channels (Edge-TTS, OpenAI-TTS, etc.), a stable internet connection is required to fetch the voice list.
   • If you are behind a proxy or firewall, configure it in Menu → Tools → Advanced Options → Network Proxy.
   • Try switching to a local channel (e.g., Qwen-TTS, Piper-TTS) to verify the issue is network‑related.

3. Reset configuration

   • Back up videotrans/params.json and delete it.
   • Restart the software – this will reset all API keys and settings to defaults, often resolving corrupted configuration issues.

4. Update pyVideoTrans

   • Your version is v3.97 (macOS ARM). The latest version is v4.03-0626, which includes many bug fixes.
   • Update by pulling the latest code: git pull and then uv sync.

If none of the above help, please provide the full log file from logs/ (the most recent YYYYMMDD.log), especially the lines around this error, for further analysis.

Please refer to related documentation:

• Dubbing channel introduction & setup
• Advanced options – Dubbing settings
• FAQ – Edge-TTS errors

I installed F5-TTS locally, run it as a host locally, and provide the URL to pyvideotrans, I cannot set it up correctly and getting the error above

F5-TTS 角色列表加载为 None，导致 AttributeError，根本原因通常是：

1. API 地址未正确填写（pyvideotrans 无法连接 F5-TTS 服务）
2. 参考音频未正确配置（F5-TTS 渠道无可用音色）

按以下步骤排查与修复：

1. 确认 F5-TTS 服务正在运行

• 在浏览器中打开 http://127.0.0.1:7860，应能看到 F5-TTS 的 WebUI 页面。
• 如果无法访问，先启动 F5-TTS 服务（参考其文档启动 webui.py 或 app.py）。
• 注意：不要填写 0.0.0.0，必须改为 127.0.0.1。如果填 0.0.0.0 会报错。

2. 在 pyvideotrans 中正确配置 API 地址

• 打开 pyvideotrans 菜单 → TTS 设置 → F5TTS（注意：F5-TTS 与 Spark/Index/Dia/VoxCPM/Confucius4 共用同一个设置面板）。
• 在 API 地址 栏填写：http://127.0.0.1:7860（默认端口，若你修改过则用实际端口）。
• 点击 保存。

3. 配置参考音频（必须）

F5-TTS 是克隆式 TTS，必须提供参考音频才能生成音色列表。

• 在软件根目录下找到 f5-tts 文件夹（若不存在则手动创建）。
• 放入一段 3~10秒 的 wav 格式 参考音频，命名为简短英文，例如 myvoice.wav。

• 打开 菜单 → TTS 设置 → 设置参考音频，在新行中添加：

  myvoice.wav#该音频中说的文本内容

• 保存并退出。

4. 选择配音角色

• 回到主界面，选择 Dubbing Channel = F5-TTS。
• 在 Select Voice 下拉框中应能看到你刚配置的 myvoice.wav，选中它。
• 如果下拉框仍然为空（或报错），说明第 2 或第 3 步未生效，请检查日志。

5. 检查日志

打开 logs/ 目录下最新的日志文件，搜索 F5-TTS 或 f5，查看是否有连接失败、文件未找到等具体错误。若仍有问题，请提供日志末尾 30 行。

参考文档

• F5-TTS 渠道使用说明
• 参考音频配置方法
• F5-TTS 官方部署