Search Mashup Node
The Search Mashup node queries the VIDIZMO content library to find videos, documents, images, and other media using comprehensive filtering and semantic search capabilities. It supports both traditional keyword-based queries and vector search for finding conceptually similar content. Search results include only the content parts specified, reducing data transfer and improving performance.
How It Works
When the node executes, it constructs a search request based on configured filters, sends it to the VIDIZMO search service, and retrieves matching content with their metadata. The node provides extensive filtering across multiple dimensions including content types (video, audio, document, image, live, collection, playlist), temporal ranges (creation dates, publication dates), content features (transcription, AI processing, featured status), and custom attributes.
Multiple filters can be combined to create precise queries, such as finding all AI-processed videos published in the last month with specific tags. Search results are returned as an array of mashup objects stored in the specified output variable, with each object containing the requested data parts. The node supports pagination through the Max Results parameter (1 to 1000 items), and when searching collections or playlists, nested mashups can optionally be included.
Variable interpolation across all parameters enables dynamic search criteria based on user input, conversation context, or previous workflow results.
Configuration Parameters
Output Field
Output Field (Text, Required): Workflow variable where search results are stored.
The output is a search result object containing Mashups (array of matching content), TotalCount (total matches found), and PageIndex (current page). Each mashup object contains the data parts specified in Mashup Search Parts.
Common naming patterns: search_results, mashup_results, content_list, filtered_mashups.
Mashup IDs
Mashup Ids (Array, Optional): Specific mashup IDs to retrieve exact content items.
Accepts an array of integer IDs or a variable reference using ${variable_name} syntax. When specified, the search returns only mashups matching these IDs, ignoring other filter criteria.
Keyword
Keyword (Text, Optional): Search terms to filter content by title or description.
Supports full-text search with automatic relevance ranking. Multiple words provide broader matches; quoted phrases provide exact matches. Variable interpolation with ${search_query} supports dynamic search terms.
Author Name
Author Name (Text, Optional): Content creator name to filter by.
Searches against the author field in content metadata. Supports partial name matching and variable interpolation with ${author}.
Tags
Tags (Array, Optional): Tags to filter content.
Accepts an array of tag strings for multi-tag filtering. Content matching any specified tag is returned. Tags are case-insensitive.
Vector
Vector (Text, Optional): Named vector for semantic search.
Must reference a variable containing an object with name (string) and vector (array of floats) fields. The vector name should match an embedding field configured in VIDIZMO (e.g., "description_embedding", "transcript_embedding"). Semantic search finds content with similar meaning rather than exact keyword matches.
Content Types
Content Types (Dropdown, Optional): Media formats to include in search results.
| Content type | Description |
|---|---|
| Video | Video content including uploaded files and recordings |
| Audio | Audio-only content and podcasts |
| Document | PDF, Word, PowerPoint, and other document formats |
| Image | Image files and photo galleries |
| Live | Live streaming sessions and webinars |
| Collection | Content collections and folders |
| Playlist | Ordered playlists of content |
Multiple types can be selected. Leave empty to search all content types.
Timed Data Types
Timed Data Types (Dropdown, Optional): Filter content containing specific types of timed data.
- Chapters - Content with chapter markers for navigation
- Video Description - Content with timed descriptions or annotations
Only returns content that has the selected timed data types.
Include Nested Mashups
Include Nested Mashups (Toggle, Default: false): Include children of collections and playlists in results.
When enabled, searches within collections and playlists to find individual content items. When disabled, only returns top-level content.
Is Featured
Is Featured (Toggle, Default: false): Filter to featured content only.
When enabled, only returns content marked as featured. Featured content is typically highlighted or promoted content in the portal.
Has Transcription
Has Transcription (Toggle, Default: false): Filter to content with captions or transcripts.
When enabled, only returns content with text alternatives for accessibility or searchable transcripts.
AI Processed
AI Processed (Toggle, Default: false): Filter to AI-enhanced content.
When enabled, only returns content processed by AI services (transcription, vision analysis, redaction, etc.).
Added From Date
Added From Date (Text, Optional): Content creation start date.
Uses ISO 8601 UTC format (YYYY-MM-DDTHH:MM:SSZ). Variable interpolation with ${start_date} is supported. Combine with Added To Date to create date ranges.
Added To Date
Added To Date (Text, Optional): Content creation end date.
Uses ISO 8601 UTC format. Variable interpolation with ${end_date} is supported.
Published From Date
Published From Date (Text, Optional): Publication start date.
Uses ISO 8601 UTC format. Publication dates differ from creation dates—content may be created but published later.
Published To Date
Published To Date (Text, Optional): Publication end date.
Uses ISO 8601 UTC format. Combine with Published From Date to create publication date ranges.
Max Results
Max Results (Number, Default: 20): Maximum number of results to return (1-1000).
Lower values provide faster responses; higher values retrieve more content per request. For large result sets, consider pagination or multiple searches.
Mashup search parts
Mashup Search Parts (Dropdown, Optional): Parts of mashup data to fetch in results.
| Part | Data Included | Use When |
|---|---|---|
| BasicInfo | Title, description, ID, format, duration, thumbnail | Core content metadata for display or filtering |
| TimedData | Chapters, transcripts, annotations, timed metadata | Processing timed content or building navigation |
| Content | File details, encoding information, source files | Analyzing content structure or processing media |
| PlaybackUrl | Streaming URLs, player configuration | Embedding or playing content in custom interfaces |
| CustomFields | Custom attribute values and metadata | Working with organization-specific metadata |
| Categories | Category assignments and taxonomy | Filtering or organizing by content categories |
| Comments | User comments and discussions | Analyzing engagement or building social features |
| Permissions | Access control and sharing settings | Checking content visibility or permissions |
Defaults to BasicInfo, TimedData, Content, and PlaybackUrl if not specified. Select only needed parts to reduce response size.
Common Parameters
This node supports common parameters shared across workflow nodes, including Stream Output Response and Logging Mode. For detailed information, see Common Parameters.
Best Practices
- Start with broad search criteria and progressively narrow using additional filters rather than overly specific queries
- For semantic search, generate embeddings from user queries using an Embedding node, then pass the vector variable
- Select only needed Mashup Search Parts to reduce response time and memory usage
- Variable interpolation makes search parameters dynamic based on user input or conversation context
- Limit initial results to 20-50 items for faster responses; implement pagination if more results are needed
- Combine filter types strategically: filter by Content Types first, then apply date ranges and feature filters
- Enable Include Nested Mashups for user-facing search to find content within collections and playlists
Limitations
- Search Scope: The node searches only content accessible to the workflow's execution context (tenant and user permissions). Inaccessible content is not returned.
- Vector Search Requirements: Semantic vector search requires pre-configured embedding fields in VIDIZMO. The vector name must match an existing embedding field.
- Max Results Limit: Maximum 1000 items per request. Larger result sets require pagination with multiple search requests.
- Date Format Requirement: All date parameters must use ISO 8601 UTC format (YYYY-MM-DDTHH:MM:SSZ). Invalid formats cause search failures.
- Timed Data Availability: Filtering by Timed Data Types only returns content that has been processed. Newly uploaded content may not have timed data until processing completes.
- Performance With Large Results: Requesting all Mashup Search Parts with Max Results set to 1000 results in large response payloads and increased processing time.