Community Solution

ZimaOS Search: Basic Index vs AI Search Explained

A 2024 user asked for Files search by extension; IceWhale said search and AI search were coming, and current ZimaOS now has a full indexing system.

ZimaOS now has a full file-search system, so the 2024 “Search is coming soon” answer is obsolete. Basic search indexes filenames and supported document content, while the optional ZimaOS-AI module adds OCR, audio/video extraction, and semantic image search when your hardware meets its requirements.

If the search box is completely missing, check Settings → General → Developer Mode → Search Index before installing AI components. Basic Search and AI Search are related but not the same feature.

What Basic ZimaOS Search Does Today

The current ZimaOS Search architecture guide describes real-time filename monitoring, content indexing, fuzzy/exact filename matching, and support for more than 70 file types.

Enable Search Index If the Search Box Is Missing

A user-verified fix on current-era ZimaOS is Settings → General → Developer Mode → Search Index. Reload Files after enabling it.

Indexing Is Not Instant for Every Content Type

Filename changes are monitored in real time, while heavier content indexing can run asynchronously and during off-peak processing. A newly copied library may therefore appear by filename before all content becomes searchable.

AI Search Is an Optional Extra Layer

The current ZimaOS AI Search guide says the AI module adds image OCR/tagging, audio/video transcription, and semantic capabilities.

You Do Not Need a GPU for Basic File Search

The old forum discussion correctly separated simple search from AI features. Filename and ordinary indexed document search are the baseline. AI processing has separate NVIDIA/Intel hardware and free-space requirements.

Search by File Extension

The current documentation emphasizes filename/content matching rather than a dedicated “extension filter” control. For a precise bulk list of one extension, SMB/Desktop search or shell tools can be more predictable:

find /path/to/folder -type f -iname '*.pdf'

Use read-only listing commands first before any bulk move/delete operation.

Why Search Can Use CPU in the Background

Indexing thousands of new files requires metadata and content extraction. Large imports can temporarily increase CPU and disk activity, especially when Search Index or AI feature extraction is first enabled.

Troubleshoot a Search That Finds Nothing

Confirm Search Index is enabled, wait for indexing, check the exact folder is part of Files, and verify the file type is supported. For AI results, also check model download and journalctl -xef -u zimaos-ai.

The system troubleshooting guide helps structure escalation.

Understand the Midnight Full Audit

Current Search documentation says the system performs a full audit around midnight to maintain eventual consistency. If one file is missed by real-time monitoring, the scheduled audit is designed to reconcile the index later.

Large Libraries Need Time to Settle

Copying hundreds of thousands of files and immediately judging Search can produce false negatives. Let filename indexing finish, then allow background content processing to catch up before rebuilding or disabling the index.

AI Search Has Higher Storage and Memory Costs

The optional AI module downloads models and stores them separately from ordinary search indexes. Check free system/storage space and GPU or Intel iGPU requirements before enabling it on a small home server.

FAQ

Does ZimaOS have search now?

Yes. Current ZimaOS has a documented multi-layer search/indexing system.

Do I need a GPU to search filenames?

No. GPU/AI resources are for AI-enhanced features, not the basic search box.

Why is the search field missing?

Check the Search Index toggle under Developer Mode, then reload Files.

Can ZimaOS search inside documents?

Yes for many supported formats; current docs describe full-text content indexing alongside filename search.