Analyzer

Find items that didn’t sync and identify missing IDs or mismatches.

Analyzer helps explain why items do or do not line up between your selected providers.

What it does

• Reads state.json from your latest runs.

• Uses the selected pair to show actual sync issues.

• Separates sync problems from background diagnostics.

• Helps you inspect weak or inconsistent IDs.

Analyzer does not “fix metadata for you”. It helps you find the right fix faster.

What you need

• At least one recent sync run (so state.json exists).

• The pair and feature must be enabled.

Optional, for better suggestions:

• TMDb key (see Metadata)

• Trakt app credentials (if you use Trakt lookups)

What the sections mean

Issues

These are sync-related problems in the selected pair, such as:

• an item exists on one side but is missing on the other

• an item is blocked by manual rules

• an item is out of scope because of pair settings

• an item has weak or inconsistent IDs that can cause matching trouble

System

These are background diagnostics about CrossWatch state and provider data, such as:

• broken or stale cache/state files

• unresolved, shadow, flap, blackbox, or watermark problems

• provider/module validation issues

• metadata inconsistencies that may not be causing an active sync problem yet

Scoped

This is the number of items included for the currently selected pair filter.

Visible

This is the number of rows currently shown in the top table after search and scope filters are applied.

Issues tell you what is wrong in the selected sync path.

System findings help explain why.

How to use it

1

Run a sync first

Analyzer works from runtime state. Do one normal sync run before you start debugging.

2

Select the pair you want to inspect

Start with the pair that looks wrong right now.

3

Look at the top table

Start with actual sync issues in the top table.

4

Click an item for details

Review the reason details in the lower panel.

5

Strengthen IDs if needed

Open Edit Manual IDs if the item needs stronger IDs.

6

Use system findings as supporting evidence

Review System findings when something still looks wrong after checking the item itself.

7

Fix at the source, then re-run

For media-server → tracker pairs:

1. Fix IDs in Plex/Jellyfin/Emby.

2. Refresh metadata if needed.

3. Re-run the pair.

4. Confirm the issue disappears.

Where to fix issues

In almost all setups, fix the source of the pair.

Media server → tracker:

• Fix IDs and matching in Plex/Jellyfin/Emby.

• Re-run the pair.

Tracker-only items:

• Analyzer does not edit trackers directly.

• Clean it up in the tracker, or import it to your media server.

Issue types (reference)

Common issue types

• missing_peer: item exists on source, but no match exists on the target for this pair.

• missing_ids: core IDs (IMDb/TMDb/TVDb) are missing.

• invalid_id_format: ID format is wrong (example: IMDb missing tt).

• key_missing_ids: key claims an ID, but the ID field is empty.

• key_ids_mismatch: key says one ID, field says another.

Suggestions are hints, not guarantees. Sanity-check before trusting them.

UI guide (quick)

• Pair selector: focus on one selected pair.

• Issues: actual sync problems for that pair.

• System: supporting diagnostics and state warnings.

• Top table: sort and search by provider, feature, title, year, and type.

• Detail panel: shows likely cause and suggestions.

• Edit Manual IDs: lets you test stronger IDs locally.

• Scoped: items included for the selected pair filter.

• Visible: rows still shown after search and scope filters.

Related: Editor.

Screenshots

Analyzer – top controls

Analyzer – grid

Analyzer – detail panel

Analyzer – manual IDs editor

Troubleshooting

• Scoped looks wrong: check the selected pair and pair settings.

• Visible looks too low: clear search or other active filters.

• No suggestions: add TMDb metadata, or connect Trakt for richer lookups.

• System findings keep appearing: inspect local state freshness, provider health, and cache behavior.

• Same issue returns after sync: source metadata is still wrong. Fix IDs/year and re-run.