Knowledge base — upload, import, and search
Upload documents, images, audio, and video
You can upload the following directly to a project or workspace knowledge base:
- PDFs, Word documents, spreadsheets, and text files;
- Images;
- Audio and video recordings.
Large files are uploaded directly to cloud storage instead of being held in your browser. After uploading, you can rename a file, move it into a folder, or delete it.
Import from Google Drive
After authorizing your Google account, you can search for and import files directly from Drive without downloading them locally and uploading them again.
Folders, privacy, and sharing
Use folders to organize knowledge base content. Each folder can have:
- A summary description;
- A privacy scope that controls who can view it;
- Sharing access for specific members or collaborators.
Transcription, webpage capture, and analysis
- Audio: Uploaded recordings are transcribed automatically. You can transcribe them again or revise the transcript. Long recordings are divided into searchable segments.
- Webpages: Paste a URL and GDDAO will analyze and save the page content so it can later be searched and cited.
- Images: Images embedded in documents can be extracted for individual use.
Search and cite knowledge in conversations
Knowledge base content becomes searchable after processing. In an AI conversation, you can ask it to read a specific project or folder—for example, “Find all data related to the number of people served by this activity.” AI only reads information within your authorized scope.
Tips
- Saving a file does not automatically trigger deep AI analysis. Standard attachments incur only basic storage usage; full parsing occurs when you explicitly need AI search or analysis.
- Uploaded files display a processing status: processing, completed, or failed. You can upload the file again if processing fails.