You can access your app settings from the top right corner of the Drive view.
General settings
| Setting | Description |
|---|---|
| Default project | Set whether new projects default to video or audio-only compositions. |
| Default AI speaker | Set a default speaker label for new Write mode projects—ideal for creating content from scratch. |
| Theme | Choose Light, Dark, or match your system settings. |
Transcription settings
| Setting | Description |
|---|---|
| Default transcription language | Set to Auto by default. Optionally choose from 23 supported languages for automatic transcription. |
| Always ask before transcribing | Prompts you to confirm the language before transcribing. Useful if you work in multiple languages. This is also useful if you do not want transcribe a file you import. By choosing exit out of picking a language, will cause the file to not be transcribed. |
| Transcription provider | Choose your preferred transcription engine. Hover over the info icon in-app to compare provider strengths. |
| Automatically detect speakers | When enabled, Descript asks for confirmation before running speaker detection. Learn more about detecting speakers. |
Advanced settings
| Setting | Description |
|---|---|
| Automatic volume levels | Applies -16 dB loudness normalization. Can be overridden per project. |
| Use room tone on gap clips | Adds room tone to gap clips. Can be overridden per project. |
| Apply Smart transitions | Automatically adds Smart transitions between scenes. |
| Show advanced timeline tools | Enables the advanced timeline tools menu. |
| Show wordless media in script | Displays waveforms for non-speech media in the script. Learn more about wordless media. |
| AI speech generation model | Choose which AI model powers text-to-speech voice generation. Learn more about generating text-to-speech with AI speakers. |
| Image generation model | Choose which model powers AI-generated images. Generate AI b-roll or create images with AI Tools. |
| Force software video decoding | Disables GPU-based video decoding and uses your CPU instead. Useful if you see blank previews, editor freezing, or playback stalls. |
Choose your generative media models
The Model Picker lets you select which AI model powers your generative media tasks. Different models balance quality, speed, and AI Credit usage, so you can pick what works best for your workflow.
Different models have different strengths—some are faster, others handle complex prompts or non-English text more accurately. Premium models use more AI Credits. Your selection only applies to new generations; existing media won’t be re-processed.
Availability of models may vary by plan. Learn more about generative image and video models in Descript
Labs
Enable early-stage features to test and share feedback.
Learn more about Labs.