> For the complete documentation index, see [llms.txt](https://docs.cherryai.com.cn/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cherryai.com.cn/docs/en-us/pre-basic/settings/yu-yin-gong-neng.md). # Voice Features {% hint style="warning" %} This feature was put on hold because the relevant developer did not continue maintaining the PR. {% endhint %} Cherry Studio Voice Feature Instructions 1\. Overview of Voice Features Cherry Studio provides three major voice feature modules: TTS (text-to-speech), ASR (automatic speech recognition), and voice calling. These features let you communicate naturally with AI by voice, improving the overall experience. * TTS (text-to-speech): Converts AI replies from text into spoken output * ASR (automatic speech recognition): Converts your speech into text input * Voice calling: Combines TTS and ASR to deliver a ChatGPT-like voice conversation experience 2\. TTS (Text-to-Speech) Features 1. Supported service types Cherry Studio supports four TTS service types: * OpenAI: Uses OpenAI's TTS API, requires an API key * Browser TTS: Uses the browser's built-in speech synthesis, free and requires no configuration * SiliconFlow: Uses SiliconFlow's TTS service, requires an API key * Free online TTS: Uses a free online TTS service, requires no API key 2. Setup method 3. Go to the Settings page and select the "Voice Features" tab 4. In the "TTS" sub-tab: * Enable TTS (turn on the switch) * Select the TTS service type * Configure the corresponding parameters according to the selected service type: * OpenAI: Enter the API key, API address, and select a voice and model * Browser TTS: Select a voice * SiliconFlow: Enter the API key, API address, and select a voice, model, response format, and speech rate * Free online TTS: Select a voice and output format 5. Configure TTS filtering options (optional): * Filter thought process * Filter Markdown markup * Filter code blocks 6. Set whether to show the TTS progress bar 7. Click the "Test TTS" button to test whether the configuration is correct 8. How to use * After TTS is enabled, AI replies will automatically be converted to spoken output * In the chat interface, a TTS playback button will appear below each AI reply * Click the play button to play/pause the speech * If the TTS progress bar is enabled, playback progress will be shown below the text * Long text will be automatically segmented and synthesized continuously 3\. ASR (Automatic Speech Recognition) Features 1. Supported service types Cherry Studio supports three ASR service types: * OpenAI: Uses OpenAI's Whisper model, requires an API key * Browser: Uses the browser's built-in speech recognition, free and requires no configuration * Local server: Connects to a local WebSocket server for speech recognition 2. Setup method 3. Go to the Settings page and select the "Voice Features" tab 4. In the "ASR" sub-tab: * Enable ASR (turn on the switch) * Select the ASR service type * Configure the corresponding parameters according to the selected service type: * OpenAI: Enter the API key, API address, and select a model * Browser: No additional configuration required * Local server: You can set whether to automatically start the ASR server when the app launches * Select the speech recognition language (Chinese by default) 5. Click the "Test ASR" button to test whether the configuration is correct 6. How to use * After ASR is enabled, a speech recognition button will appear next to the input box * Click the speech recognition button to start recording * After you speak, the speech will be converted to text and filled into the input box * Click the button again to stop recording * Speech recognition supports continuous recognition of multiple sentences using an accumulation mode 4\. Voice Calling Features 1. Feature highlights * Combines TTS and ASR to deliver a ChatGPT-like voice conversation experience * Uses a draggable floating window interface * Supports press-and-hold-to-speak mode * Supports custom shortcuts * Supports window collapsing * You can choose a dedicated voice-calling model * Supports custom prompt words 2. Setup method 3. Go to the Settings page and select the "Voice Features" tab 4. In the "Call Features" sub-tab: * Enable voice calling (turn on the switch) * Click the "Choose Model" button to select the AI model used for voice calling * Customize the voice calling prompt text in the prompt text box (optional) * Click the "Save" button to save the prompt, or click "Reset" to restore the default prompt 5. How to use 6. In the chat interface, click the voice calling button (phone icon) to the right of the input box 7. The voice calling window will open and play a welcome message 8. Press and hold the "Press and Hold to Speak" button to start recording (or use the shortcut you set) 9. Release the button to end recording and send it to AI for processing 10. AI generates a reply and plays it through TTS 11. Use the control buttons in the window: * Mute/Unmute button: Controls TTS output * Pause/Resume button: Pauses or resumes the conversation * Settings button: Configure shortcuts * Collapse button: Collapse the window, leaving only the press-and-hold-to-speak row 12. Click the close button to end the call 13. Keyboard shortcut settings 14. In the voice calling window, click the Settings button 15. In the settings panel that appears, click the shortcut button 16. Press the key you want to set (such as the Space key, Shift key, etc.) 17. Click the "Save" button to save the settings 18. When using it, press and hold the configured shortcut to start recording, and release to end recording and send 5\. Common Problems and Solutions 1. TTS-related issues * Problem: TTS cannot play audio. Solution: Check whether TTS is enabled, and make sure the correct service type is selected and the required parameters are configured. * Problem: Poor TTS playback quality. Solution: Try switching to a different TTS service type or voice. * Problem: TTS shows an error message during playback. Solution: Check whether the API key is correct and whether the network connection is normal. 2. ASR-related issues * Problem: ASR cannot recognize speech. Solution: Check whether ASR is enabled, and make sure the correct service type is selected and the required parameters are configured. * Problem: ASR recognition accuracy is low. Solution: Try switching to a different ASR service type, or adjust the microphone position and volume. * Problem: ASR server connection failed. Solution: Check whether the local server is running properly, or try restarting the app. 3. Voice calling-related issues * Problem: The voice calling window cannot be opened. Solution: Check whether voice calling is enabled, and make sure the TTS and ASR features are configured correctly. * Problem: Press-and-hold to speak does not respond. Solution: Check whether microphone permission has been granted, or try restarting voice calling. * Problem: AI replies have no audio output. Solution: Check whether TTS is enabled and make sure it is not muted. 6\. Advanced Settings and Custom Options 1. TTS advanced settings * Filtering options: You can choose to filter the thought process, Markdown markup, and code blocks to make TTS playback smoother * Progress bar display: You can choose whether to show the TTS progress bar * Custom voices and models: You can add custom voice and model options 2. ASR advanced settings * Auto-start server: You can set whether the ASR server automatically starts when the app launches * Language selection: You can choose different speech recognition languages 3. Voice calling advanced settings * Custom prompts: You can customize the voice calling prompt to guide how AI responds in voice calling mode * Dedicated model selection: You can choose a dedicated AI model for voice calling, separate from the model used in the current conversation * Shortcut customization: You can set custom shortcuts to control recording 7\. Usage Recommendations 1. Choose the appropriate TTS service: * If you are pursuing high-quality speech, OpenAI or SiliconFlow is recommended * If you do not want to configure an API, you can use Browser TTS or Free online TTS 2. Choose the appropriate ASR service: * If you are pursuing high accuracy, OpenAI is recommended * If you do not want to configure an API, you can use the browser's built-in speech recognition 3. Optimize the voice calling experience: * Using headphones can prevent TTS output from being captured again by ASR * Using it in a quiet environment can improve recognition accuracy * Using custom prompts can make AI replies more suitable for speech playback 4. Adjust settings according to your needs: * If you mainly use text communication, you can enable only the TTS feature * If you mainly use voice input, you can enable only the ASR feature * If you need a complete voice conversation experience, enable the voice calling feature We hope this user guide helps you make full use of Cherry Studio's voice features and enjoy a more natural and convenient AI interaction experience! *** ### 💡 Get help and submit feedback If you encounter any questions, bugs, or have suggestions for feature improvements during configuration or use, please refer to [Feedback and Suggestions](/docs/en-us/question-contact/suggestions.md) the official channels provided there. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.cherryai.com.cn/docs/en-us/pre-basic/settings/yu-yin-gong-neng.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.