Skip to content

Files & Multimodal Processing

Files & Multimodal Processing

AMC WebUI provides robust pipelines for parsing project repositories, compressing media, and managing cloud storage assets.


1. ZIP & Folder Codebase Parsing

  • Drag & Drop: Drop any project folder or .zip archive directly into the chat input.
  • Smart Directory Tree: Automatically filters out noise directories (node_modules, .git, dist), parses file hierarchies, and constructs unified multi-file prompts for codebase refactoring and analysis.

2. Client-Side Audio Compression Worker

Audio inputs can quickly consume massive token allocations and bandwidth:

  • AMC WebUI features a dedicated browser-based Audio Compression Worker.
  • Recorded or uploaded audio is resampled, silence-trimmed, and compressed client-side, reducing up to 70% of audio token usage while maintaining high recognition fidelity.

3. Gemini Cloud Files API Manager

For large documents, extended video clips, or high-res imagery:

  • Policy Selection: Toggle between inline Base64 payloads and Google’s hosted Cloud Files API.
  • Cloud Files Modal:
    • View all active cloud files tied to your API key, including byte sizes, creation dates, and expiration countdowns.
    • Delete expired or obsolete assets with a single click.

4. Document & Media Previews

  • PDF Viewer: Embedded pdfjs-dist viewer with multi-page navigation and text selection.
  • Office Previews: Integrated docx-preview and xlsx (SheetJS) render Word documents and Excel spreadsheets into responsive tables.
  • Visual Detail Levels: Configure visual fidelity (Low, Medium, High, Ultra High) for precision engineering diagrams and fine OCR tasks.