Library Intelligence: Catalogue, Storage and Engagement Evidence
Use Finetic Library Intelligence to inspect catalogue growth, storage distribution, media formats, health, engagement and conversion demand from retained installation evidence.
1Catalogue evidence2Growth and formatsWhat Library Intelligence measures#
Library Intelligence turns the catalogue, probe results, scan history and retained playback history on one Finetic installation into an operational view. It does not inspect another customer's server and it does not send media titles to the public website.
The page has four views:
| View | Question it answers |
|---|---|
| Overview | How large, healthy and technically diverse is the catalogue? |
| Needs attention | Which missing, unavailable, ambiguous or conversion-heavy items need review? |
| Engagement | Which retained titles are watched, completed or never played? |
| Conversion load | Which sources repeatedly require remuxing or conversion? |
Evidence scope#
Every value is calculated from records retained by the selected Finetic installation. The page does not invent disk capacity, compare against an unstated industry average or estimate data that the server has not measured.
The date and catalogue filters apply to the evidence shown. A title filter narrows catalogue-derived measures to matching media. Library selection narrows media to the selected configured root.
Overview cards#
| Measure | Exact meaning |
|---|---|
| Total titles | Active catalogue records in the current scope |
| Indexed storage | Sum of known source-file sizes, not total filesystem usage |
| New titles | Active catalogue items first discovered during the last seven days |
| Never played | Scoped titles with no retained playback-history record |
| Conversion hours | Retained watched time delivered through a method other than Direct Play or HLS remux |
| Collections | Distinct retained collection names containing at least one scoped title |
| Library health | A bounded catalogue indicator derived from unavailable/error items and duplicate-candidate groups |
Library health is not a PostgreSQL health check, filesystem SMART result or promise that every title will play. Those belong to System health and playback evidence.
Growth and storage#
Catalogue growth groups title discovery dates into daily counts for the latest 30 days. It reports when Finetic first retained catalogue identity; it is not a graph of file creation dates.
Storage by library groups known source sizes by configured media root. Files without a known size do not contribute bytes. If a source is mounted through Docker, the displayed root is the path visible inside the Finetic container.
Resolution, video, audio and subtitles#
Format cards are built from the latest stored probe for each scoped source:
- Resolution uses probed width and height.
- Video codecs use the primary video stream.
- Audio codecs use the primary audio stream.
- Subtitle formats group text, image-based and other retained subtitle streams.
An Unprobed or Unknown group is evidence that Finetic does not yet have the relevant fact. It is not silently folded into another codec.
Catalogue health evidence#
The health panel exposes the inputs behind its headline:
- Probed titles
- Completed library scans
- Unavailable or error items
- Duplicate-candidate groups
Use Needs attention to inspect the actual records. Finetic keeps missing storage visible so an unavailable mount is not mistaken for deliberate deletion.
Engagement#
Engagement is based on retained canonical watch history:
- Played titles have at least one history session.
- Completed titles have at least one completed history session.
- Never played titles have no retained history session in scope.
- Watch hours use actual retained watched seconds, not the furthest seek position.
Retention settings affect these measures. Removing old history deliberately changes the available evidence.
Conversion load#
Conversion load groups retained playback by source and delivery method. It helps identify titles repeatedly requiring:
- Container remux
- Audio-only conversion
- Full video conversion
This view describes what actually happened in retained history. It does not override the playback planner or claim that the same route will be selected on every device.
Related administration pages#
- Library analyser, jobs and scheduled tasks explains how the underlying catalogue facts are produced.
- Analytics covers playback sessions, devices, bandwidth and performance.
- Playback planning explains route selection and compatibility learning.
- Diagnostics and support explains how to share evidence safely.