Media Types Guide

Introduction

The Wickson API supports a wide range of media types, enabling comprehensive semantic understanding and vector-based search across diverse content formats. This guide details the supported file types, processing capabilities, size limitations, and optimization strategies for each media category.

Supported Media Categories

The API processes five primary media categories:

Media Type	Description	Use Cases
Documents	Text-based files including papers, reports, and web content	Knowledge bases, research archives, content management
Images	Visual media ranging from photos to diagrams	Product catalogs, visual archives, design libraries
Video	Motion picture content with both visual and audio components	Training materials, presentations, archival footage
Audio	Sound recordings including speech, music, and ambient audio	Podcasts, interviews, meeting recordings
3D Models	Three-dimensional object representations	Product design, architecture, engineering assets

File Format Support

Document Formats

Format	Extensions	Notes
PDF	.pdf	Preferred for complex layouts and multi-page documents
Word Documents	.doc, .docx	Fully supported with formatting preservation
Rich Text Format	.rtf	Well-supported with basic formatting
OpenDocument Text	.odt	Full support for open document standard
Plain Text	.txt	Ideal for simple text content
Markdown	.md	Preserves basic formatting elements

Web & Data Formats

Format	Extensions	Notes
HTML	.html, .htm	Extracts content while preserving basic structure
XML	.xml	Extracts text with semantic structure awareness
JSON	.json	Processes structured data with context preservation

E-Book & Email Formats

Format	Extensions	Notes
EPUB	.epub	Full support for e-book content and structure
Email	.eml	Extracts message content and metadata

Presentation & Spreadsheet Formats

Format	Extensions	Notes
PowerPoint	.pptx	Processes slides with text and visual elements
OpenDocument Presentation	.odp	Alternative to PowerPoint, fully supported
CSV	.csv	Handles tabular data effectively
TSV	.tsv	Alternative delimiter-based tabular format
Excel	.xlsx, .xls	Processes complex spreadsheets with formulas
OpenDocument Spreadsheet	.ods	Open standard spreadsheet format

Image Formats

Format	Extensions	Notes
JPEG	.jpg, .jpeg	Excellent for photographs and complex images
PNG	.png	Ideal for screenshots, diagrams, and graphics
GIF	.gif	Supports static and animated images
BMP	.bmp	Uncompressed format with lossless quality
TIFF	.tiff, .tif	High-quality format often used for scanning
WebP	.webp	Modern format with good compression
SVG	.svg	Vector format ideal for diagrams and icons
HEIC	.heic, .heif	Apple's high-efficiency image format

Video Formats

Format	Extensions	Notes
MP4	.mp4	Recommended format with excellent compatibility
QuickTime	.mov	Full support for Apple's video format
AVI	.avi	Windows-native video container format
WebM	.webm	Open web video format
MPEG	.mpeg, .mpg	Standard video format with wide support
Windows Media Video	.wmv	Microsoft's proprietary video format

Audio Formats

Format	Extensions	Notes
MP3	.mp3	Excellent for speech and general audio
WAV	.wav	Lossless audio with high quality
M4A	.m4a	Good quality with efficient compression
FLAC	.flac	Lossless compression with excellent quality
OGG	.ogg	Open format with good compression
AAC	.aac	High-quality compressed audio
AIFF	.aiff, .aif	Apple's uncompressed audio format
Windows Media Audio	.wma	Microsoft's audio format

3D Model Formats

Format	Extensions	Notes
STL	.stl	Common format for 3D printing
OBJ	.obj	Widely supported 3D format with textures
FBX	.fbx	Industry standard with limited support
GLB	.glb	Binary glTF format, highly recommended
3MF	.3mf	Modern 3D manufacturing format
PLY	.ply	Polygon file format for 3D scanning
COLLADA	.dae	XML-based format for 3D assets

File Size and Content Limits

The following limits apply to files processed by the Wickson API:

Media Type	Maximum Size	Content Limitations
Documents	20MB	128,000 tokens maximum
Images	20MB	Extremely large resolutions may be downsampled
Video	100MB	10 minutes maximum duration
Audio	50MB	20 minutes maximum duration
3D Models	50MB	Very complex models may be simplified

Note: These limits may be adjusted during the beta phase as we optimize our processing pipelines. For the most current information, check the API status endpoint.

Media Type Processing Capabilities

Document Processing

Documents undergo comprehensive processing to extract:

• Full text content with preservation of structure
• Semantic understanding of topics, themes, and concepts
• Entity recognition for people, organizations, locations, and more
• Quality analysis evaluating clarity, completeness, and relevance
• Structural analysis identifying sections, headings, and functional parts
• Relationship mapping connecting concepts within the document

Particularly effective for:

• Research papers
• Technical documentation
• Legal documents
• Reports and analyses
• Articles and publications

Image Processing

Images are analyzed to understand:

• Visual content identification of objects, scenes, and elements
• Text extraction from embedded text in images
• Scene comprehension understanding context and relationships
• Compositional analysis examining layout and design elements
• Quality assessment evaluating clarity and visual characteristics
• Aesthetic evaluation assessing visual impact and style

Particularly effective for:

• Photographs
• Diagrams and charts
• Screenshots
• Design assets
• Product imagery

Video Processing

Video content undergoes multi-modal analysis:

• Visual scene analysis frame-by-frame understanding
• Audio transcription conversion of speech to searchable text
• Timeline comprehension understanding sequence and narrative
• Key frame identification detecting significant visual moments
• Entity and action recognition identifying who and what appears
• Spoken content analysis understanding discussion topics

Particularly effective for:

• Presentations and talks
• Instructional content
• Interviews
• Product demonstrations
• Informational videos

Audio Processing

Audio files receive specialized processing:

• Speech recognition accurate transcription of spoken content
• Speaker identification distinguishing between different voices
• Non-speech audio analysis understanding music, sounds, and effects
• Topic recognition identifying discussion subjects
• Sentiment analysis detecting emotional tone
• Quality assessment evaluating clarity and listenability

Particularly effective for:

• Podcasts
• Interviews
• Lectures and speeches
• Meeting recordings
• Voice notes

3D Model Processing

3D models undergo spatial and structural analysis:

• Component identification recognizing parts and structures
• Spatial relationship mapping understanding positioning and scale
• Structural analysis evaluating design and engineering aspects
• Material classification identifying surface properties
• Purpose and function recognition understanding intended use
• Quality assessment evaluating model integrity and detail

Particularly effective for:

• Product designs
• Architectural models
• Engineering components
• Character models
• Scientific visualizations

Best Practices for File Preparation

Document Optimization

• Use PDF for complex documents to preserve formatting and structure
• Enable text layers in scanned documents for better text extraction
• Include proper headings and structure for improved comprehension
• Ensure reasonable font sizes (10-12pt for body text)
• Minimize embedded binary content that adds size without value
• Use standard fonts to ensure proper character recognition
• Balance image quality and file size for embedded graphics

Image Optimization

• Provide adequate resolution (minimum 800px on longest side recommended)
• Balance compression and quality (JPEG quality 80-90% recommended)
• Use PNG for text-heavy images like screenshots and diagrams
• Consider aspect ratio (extremely wide or tall images may process suboptimally)
• Remove unnecessary metadata to reduce file size
• Avoid heavy filters or effects that might obscure content
• Ensure proper lighting and contrast for better object recognition

Video Optimization

• Use MP4 with H.264 encoding for best compatibility
• Balance resolution and file size (720p or 1080p recommended)
• Ensure clear audio track for better transcription
• Keep videos under 5 minutes for optimal processing
• Use steady camera work for better visual analysis
• Reduce background noise in audio track
• Consider good lighting conditions for better visual recognition

Audio Optimization

• Use MP3 (128kbps+) or WAV for good quality and compatibility
• Ensure clear speech without excessive background noise
• Record at appropriate levels without clipping or distortion
• Use external microphones when possible for better quality
• Consider acoustic environment to minimize echo
• Normalize audio levels for consistent volume
• Split long recordings into logical segments

3D Model Optimization

• Use GLB format when possible for best compatibility
• Optimize polygon count while preserving important details
• Include proper materials and textures for better analysis
• Check for mesh errors like non-manifold geometry
• Use reasonable scale relative to real-world dimensions
• Include logical component naming for better identification
• Reduce unnecessary complexity in areas of low importance

Troubleshooting Media Processing

Common Issues and Solutions

Documents

Issue	Solution
Poor text extraction	Ensure document has text layer (not just images of text)
Missing structure	Use proper headings and formatting in source document
Size limit exceeded	Compress PDF or split into multiple documents
Processing errors with specific pages	Check for unusual formatting or embedded objects

Images

Issue	Solution
Poor object recognition	Improve image clarity, lighting, and contrast
Text not recognized	Ensure text is clearly visible with good contrast
Scene misinterpretation	Provide clearer images with less ambiguity
Size limit exceeded	Reduce resolution or compression quality

Video

Issue	Solution
Poor transcription	Improve audio quality and reduce background noise
Missed visual elements	Ensure adequate lighting and camera stability
Processing timeout	Reduce video length or split into segments
Size limit exceeded	Reduce resolution or use more efficient encoding

Audio

Issue	Solution
Transcription errors	Improve recording quality and reduce background noise
Speaker misidentification	Ensure clear transitions between speakers
Content misinterpretation	Speak clearly and provide context
Size limit exceeded	Use more efficient encoding or split recording

3D Models

Issue	Solution
Component misidentification	Use logical component organization
Processing timeout	Reduce model complexity
Missing details	Ensure important features have adequate resolution
Size limit exceeded	Optimize mesh and texture resolution

Advanced Media Processing Tips

Batch Processing Considerations

• Group similar media types for more efficient processing
• Use consistent file formats within batches
• Balance batch size (5-20 items recommended)
• Consider processing order for interdependent content
• Monitor batch status for large operations

Multi-Modal Content Strategies

• Leverage cross-modal connections by uploading related content
• Create logical collections for content that belongs together
• Use consistent naming conventions across media types
• Consider processing order (process reference documents before derived content)
• Balance media type distribution for comprehensive coverage

Special Use Cases

Academic Research Archives

• Prioritize document quality for research papers
• Include diagrams and figures as separate image files
• Use consistent collection organization by research area
• Consider hierarchical naming for easy navigation

Product Catalogs

• Include multiple image angles for each product
• Pair products with detailed specification documents
• Organize by product category in separate collections
• Consider 3D models for complex products

Training Materials

• Include videos with clear narration
• Pair with supplementary documents
• Organize in logical learning sequences
• Use consistent formatting across materials

---