Left Panel Settings

Creation Drawer Overview

This section documents the exact labels and toggles currently visible in the creation panel. Open the left creation drawer, then expand the relevant accordion or control group.

TTS Provider & Voice Selector

In Write Prompt mode, select your preferred speech synthesis engine under TTS provider:

• OpenAI TTS: High-speed, natural synthesized speech. Voices include Alloy, Ash, Ballad, Coral, Echo, Fable, Nova, Onyx, Sage, Shimmer, Verse, Marin, and Cedar, each labeled with character descriptions such as 'Neutral / balanced' or 'Energetic, youthful'.
• ElevenLabs (your key): Connect your custom ElevenLabs API key and voice models. Click ✎ Edit to configure keys and voice lists. If provider authentication fails or a voice is unavailable, generation stops safely and your credit is refunded.

🌐 Web Search (Fresh Info Research)

The 🌐 Web search checkbox controls real-time web research before directing text prompt scenes:

• ON - checks if your prompt needs fresh info, then researches it: The AI director checks if your prompt refers to breaking news, changing APIs, recent benchmarks, or time-sensitive data, and conducts live research first.
• OFF - generate from the prompt alone: Generates narration and scene plans directly from your authored text without external lookups.

Video Format & Layout Modals

In Upload Video mode, select how creator footage interacts with motion graphics:

• Graphics only: Source video audio is used as narration, but the video track is hidden.
• Video overlay: Creator video is pinned as a synchronized picture-in-picture overlay.
• Video + B-Roll: Alternates between talking-head presenter scenes and full-screen graphic B-Roll with animated transitions and sound effects.
• Split-screen: Creator video fills the bottom half and graphics occupy the top half (Shorts 9:16 only).

Output Segmentation: Auto vs Max Control

For uploaded audio and video files, the Output: control allows you to govern pacing:

• Auto: Uses default Whisper sentence boundaries for organic speech segmentation.
• Max Control: Lets you explicitly choose how many scenes/segments to create. The Scenes (segments) selector shows the chosen count and estimated duration per scene (e.g. 12 scenes (~5.0s each)).

The Video Settings Accordion

The collapsed Video settings accordion is accessible across all input modes and contains these exact controls:

• 🚫 Disable ASCII-(only shapes, emojis & typography): Prevents the director from generating terminal ASCII art while retaining shapes, emojis, and typography.
• ✨ Premium graphics: Enables gradients, shadows, and organic background blobs. Displays ON - gradients, shadows, blobs · brand kit disabled when active, and OFF - flat colors · brand kit applies when inactive. Note: Premium graphics and Brand Kit cannot be active simultaneously.
• 🔊 Sound effects (SFX): When ON, the LLM assigns contextual audio cues to elements; when OFF, no auto-SFX are generated.
• 🎞 Auto transitions: When ON, scenes receive animated cut-in transitions; when OFF, scenes use direct cuts.
• Economy mode: When ON, scenes may merge for lower cost and fewer scenes; when OFF, each speech beat receives its own distinct visual scene.

BG Theme & Additional Settings

Use BG Theme: to choose 🌙 Dark or ☀️ Light baseline background, text, and shape palette behavior.

The Additional settings accordion appears for Upload Audio and Upload Video modes (hidden in Write Prompt):

• Additional prompt (optional): Add positive direction (e.g. Break the first 10 seconds into 3 scenes, at least 12 scenes, fast pacing).
• Negative prompt (optional): Tell the director what to avoid (e.g. No emojis, no flashing lights, no dark backgrounds).
• Model: Exists in the application codebase but is hidden in the production UI as active models are configured by the deployment.