The TitanOS ingestion engine is designed to transform partner data into a high-performance discovery experience. This page outlines the stages your metadata goes through, from the initial feed intake to the final UI rendering.Documentation Index
Fetch the complete documentation index at: https://docs.titanos.tv/llms.txt
Use this file to discover all available pages before exploring further.
Feed Delivery & Onboarding
The speed of integration depends on the feed structure provided:- Fast-Track Integration: Partners who deliver metadata using the TitanOS Standard Schema (following our example JSONs) can bypass custom mapping phases, leading to a significantly faster “Go-Live” date.
- Custom Integration: If a partner provides a proprietary feed format, our engineering team will perform a custom mapping. Please note that this process requires additional development and testing time.
Technical Integration and Data Processing
Import
The system fetches the feed from the partner’s provided endpoint. We perform an initial validation check to ensure the file is reachable and the structure is readable.
Normalization & Mapping
Data is translated from the source format into the internal Titan OS Schema. This ensures all content, regardless of the source, speaks the same technical language within our ecosystem.
Deduplication & Aggregation
To provide a clean user experience, we ensure each unique title appears only once. We identify duplicate content using two methods:
- ID Matching: Direct matching via industry-standard IDs (IMDB or TMDB).
- Probability Scoring: If IDs are missing, our algorithm calculates a “match score” based on Title, Release Year, Runtime, Cast, and Synopsis.
Image Processing
TitanOS downloads all provided artwork URLs. Images are processed, optimized for TV performance, and cached on our CDN to ensure instant loading for the end-user.
Metadata Enhancement
We partner with TMDB to enrich your feed. This stage for example identifies similar content, and flags trending titles specific to each local market and language.
Recommendation Logic {Link to Recommendation Feed Page}
Content is organized into rows based on one of two logic paths:
- Partner-Managed: The partner retains full control over the recommendation logic via their own feed.
- TitanOS-Optimized: Our proprietary algorithm automatically generates rows based on user engagement and platform trends.
Update Frequency
- VOD Content: Processed every 24 hours during nighttime (CET) to minimize impact on system performance.
- Sports & Live Events: Processed every 30 minutes to ensure real-time accuracy for schedules and scores.