Supported files
Learn how common documents, code, data, images, audio, and video are previewed and searched.
CogLake stores the original file, derives searchable text where possible, and keeps processing failures visible to administrators.
| File family | In-app experience | Search processing |
|---|---|---|
| Markdown and text | Rich text or code editor for writable internal files | Direct text extraction and chunking |
| Embedded page preview | Text layer first; OCR can handle image-only pages when configured | |
| DOCX, ODT, PPTX, ODP | Rendered document or slide preview with extracted-content fallback | Office conversion and structured text extraction |
| XLSX, ODS, CSV | Spreadsheet or table-oriented preview | Cell and table text extraction |
| Source code, JSON, XML, HTML, SQL | Syntax-highlighted editor for writable internal files | Text extraction with file metadata |
| PNG, JPEG, WebP, TIFF | Original image preview | OCR and optional visual-language processing when configured |
| MP3 and other supported audio | Audio player | Speech-to-text when configured |
| MP4 and other supported video | Video player | Audio extraction followed by speech-to-text |
| Mail-style sender, recipient, date, subject, and body view | Message body and metadata are searchable |
Fallbacks
If a rich preview cannot be generated, CogLake shows extracted content or a clear unavailable state and keeps Open original available when permitted. Unsupported, encrypted, unsafe, oversized, or malware-positive files are classified instead of being silently dropped.
Internal versus connected files
Files in an internal workspace can be edited when your resource scope permits it. Files coming from connectors remain read-only in CogLake; edit those at the source and let the connector synchronize the change.