DocFetcher Problems & Issues
Real user-reported problems, complaints, and limitations for DocFetcher. Based on analysis of 12 signals.
Overall score: 64/100
Issues by Category
DocFetcher not finding words that are in the file
Searches match whole tokens, so part of a word or long number strings often return nothing unless you use wildcards. Scanned PDFs without a text layer, unsupported extensions and stale indexes also cause missed hits.
Scanned PDFs and images are not searchable
DocFetcher reads the text layer only. Image-only PDFs show an empty preview and never match, so they need OCR in another tool first.
Plain-text previews and limited interface
Previews are text only (HTML aside), there is no hit-count column, and folder names are not searchable in the free version. Users describe the GUI as limited compared with newer tools.
Not a filename finder like Everything
People expecting instant filename search find DocFetcher slow and heavy. It only searches folders you index, which is the point of a content search tool but surprises users coming from Everything.
Runs out of memory on large document collections
Large indexes hang or crash until the Java memory limit (-Xmx) is raised by editing a launcher file. Users on AlternativeTo describe needing bigger heap settings or disabling previews to keep it stable.
Why is DocFetcher indexing so slow?
The first index has to open and extract text from every file, which users say can take hours or even days on big collections. There is no progress bar because counting files in advance would slow it down further.
High CPU usage from folder watching and the daemon
When another program keeps changing files in an indexed folder, folder watching keeps DocFetcher busy. On Linux the daemon can sit at 100% CPU if the .indexes.txt file is missing.
Fiddly setup: config files instead of settings
Raising memory, fixing deep-folder errors, changing language or index location all mean editing .bat, .sh, .ini or .txt files by hand. Reviewers say getting it working well takes trial and error.
Older versions need a separate Java install
Up to 1.1.25 you had to install Java yourself (64-bit, and x64 Java even on Apple Silicon Macs). From 1.1.26 a runtime is bundled, but old portable setups still throw 'requires a Java Runtime' errors.
DocFetcher crashes with 'program died' while indexing
The bug tracker is full of 'program died' and Java exception reports, many triggered while parsing odd Office, PDF or image files. The FAQ says most indexing errors come from third-party parsers (Apache POI, PDFBox) and that a crashed index should be rebuilt.
Free version gets fixes, not features
The FAQ says feature requests are not planned for DocFetcher and development continues in the paid DocFetcher Pro. The SourceForge bug tracker has hundreds of open tickets, and a long-time user called it unmaintained.
Outlook won't start after indexing PST mail
The FAQ documents a Windows conflict where the DocFetcher daemon holds Outlook data files so Outlook refuses to open. The fix is renaming the daemon executable and updating indexes manually.
Regret Patterns
Common reasons users regret choosing DocFetcher.
Days of indexing, then a crash
Users point it at an entire drive, wait hours, and the index dies on one bad file or memory limit, forcing a full rebuild.
Scanned archive turns out unsearchable
People expecting it to find text in scanned contracts or receipts discover it has no OCR.
Waiting for features that only ship in Pro
Users requesting improvements learn the free edition is in maintenance mode and new work goes into the paid Pro.
Config file editing for every fix
Memory, deep folders, language and index location all require hand-editing launcher files, which non-technical users give up on.
Expected Everything-style instant results
Users wanting fast filename search find DocFetcher heavy and slow for that job and switch back.
When DocFetcher Breaks Down
Specific scenarios where DocFetcher fails or underperforms.
Indexing 100 GB+ in one go
Default memory runs out and the index can crash; the FAQ advises max memory, queuing subfolders separately and rebuilding after any crash.
Deeply nested folder trees
Indexing stops with 'Folder hierarchy is too deep' until the Java stack size (-Xss) is raised in the launcher.
Folders another program keeps changing
Folder watching keeps re-triggering, pushing CPU high; turning off watching for that index fixes it.
Indexing Outlook PST on Windows
The daemon can stop Outlook from starting until its executable is renamed.
External drives remounted under a new letter
Search still works but files no longer open or preview until the original drive letter is restored.
Searching codes, part numbers or partial words
Tokenization splits on symbols and whole words, so partial strings and IDs are missed without wildcards or the source-code analyzer.