Upload
To create a new signal, upload a file using drag-and-drop or the file picker. Supported formats include:
- Video -- MP4, MOV, WebM, AVI, MKV
- Audio -- MP3, WAV, FLAC, M4A, OGG, AAC
- Photo -- JPEG, PNG, WebP, GIF, TIFF
- Text -- Plain text, or enter text directly in the editor
Signal Type Auto-Detection
The signal type is automatically detected based on the uploaded file's MIME type. A video file becomes a VIDEO signal, an image becomes a PHOTO signal, and so on. You can also create TEXT signals by writing directly without uploading a file.
Metadata Fields
Each signal includes the following metadata:
- Title -- Auto-generated from the file name or content, but can be edited
- Date -- The date the signal was captured. For photos and videos with EXIF data, this is extracted automatically. Otherwise, it defaults to the upload date and can be set manually
- Location -- GPS coordinates, pulled from EXIF data when available. Can also be set manually by placing a pin on the map
- Visibility -- Choose between Public, Sanctum, or Private at creation time
What Happens After Upload
Once a signal is created, the processing pipeline kicks in automatically:
- The file is transferred to permanent storage
- A thumbnail is generated (for video and photo signals)
- Transcription begins (for video and audio signals)
- AI analysis runs in two passes: surface-level observations followed by structural analysis
- A vector embedding is generated for semantic search
No manual intervention is required. The entire pipeline is automated.
Processing Status
The notification panel displays real-time progress for each processing step. You can see which step is currently running, whether any step has failed, and when processing is complete. Once all steps finish, the signal is fully indexed and searchable.