Latest News

AI File Sorter 1.6.0 is now available

AI File Sorter 1.6.0 adds a major new capability: document content analysis using a text LLM (not just filenames/extensions), plus smarter reliability, review workflow upgrades, and broader model/connectivity options.

Document content analysis and rename suggestions

AI File Sorter can now read and analyze document text and use it to suggest categories and human-readable filenames. For example, a generic name like scan_0021.pdf can be renamed to something descriptive such as 2024_tax_residency_letter.pdf. Supported formats include PDF, DOCX, XLSX, PPTX, ODT, ODS, and ODP, plus common text formats. You can also use rename-only or document-only modes, and optionally add creation-date suffixes to category names. All rename suggestions are optional and must be reviewed and approved by you before being applied.

More flexible LLM selection and custom endpoints

The Select LLM dialog is faster, and now supports custom API endpoints. You can enter either a base URL or a full /chat/completions endpoint, with optional API keys for local servers. A new system compatibility check (benchmarking) helps pick the most suitable model for your machine.

Better stability: progress is saved as you go

LLM-derived categorizations and rename suggestions are now persisted continuously, so you don’t lose progress if the app closes unexpectedly.

Improved GPU handling for image analysis

If there isn’t enough available GPU memory during image analysis, AI File Sorter can now prompt you to fall back to CPU so the run can continue instead of failing. The Local 3B download now defaults to Q4 for better GPU compatibility, while the legacy Q8 variant remains available when an existing download is detected.

Review dialog workflow upgrades

The review dialog now supports bulk category edits for highlighted rows, and it’s scrollable on smaller screens so action buttons stay visible. Subcategory labels are also more consistent thanks to automatic merging of near-duplicates that only differ by generic suffixes (for example, labels ending in “files”).

Localization and platform updates

Korean has been added as a selectable interface language. On macOS, builds now include separate make targets for Apple Silicon (M1 / M2–M3) and Intel, with improved arch-aware llama.cpp builds.

AI File Sorter 1.5.0 is now available

AI File Sorter 1.5.0 brings a major new capability: local image content analysis with intelligent filename suggestions, along with workflow and usability improvements.

Local image content analysis and rename suggestions

AI File Sorter can now analyze picture files directly on your device and suggest meaningful, human-readable file names. For example, a generic name like IMG_2048.jpg can be renamed to something descriptive such as clouds_over_lake.jpg. All rename suggestions are optional and must be reviewed and approved by you before being applied.

Improved review dialog

The review dialog has been enhanced to support rename-only workflows, inline editing of suggested filenames, and clear status labels indicating whether files will be Renamed or Renamed & Moved. This makes it easier to understand and control exactly what will happen before confirming any changes.

Picture-only processing mode

A new picture-only processing toggle lets you focus runs exclusively on supported image files when needed.

Localization updates

Dutch has been added as a selectable interface language, and the analysis progress dialog output is now localized across all supported UI languages.

Privacy-first by design

AI File Sorter can run entirely on your device using local AI models. No files, filenames, images, or metadata are uploaded, and no telemetry is sent anywhere. An internet connection is only used if you explicitly choose to enable a remote model.

This update brings a preview (dry-run) mode, persistent Undo, and a few UI and safety improvements. The remote LLM setup now uses your own OpenAI API key, with support for any ChatGPT model. Simple, clean, no fluff.

You can now add your own Local LLM (custom .gguf models in the Select LLM dialog); the app also has two categorization modes - More Refined + More Consistent; Optional Whitelists allow restricting category names according to your use case. Multilingual categorization & more UI languages (Dutch, French, German, Italian, Polish, Portuguese, Spanish, Turkish). And more...

This version brings major improvements and paves the way toward the 1.0 milestone:

• Local LLM Support – You can now run file categorization without an internet connection using a local Mistral or LLaMa model.
• CUDA, OpenCL, and OpenBLAS support for faster AI compute.
• New Onboarding Guide – First-time setup now helps you choose and download a model.

📂 Less clutter. More clarity.

This is the latest version of AI File Sorter, an intelligent tool for ordering your files and directories using AI. Some code refactoring has been done in order to improve the program design and code readability. Also, hidden files are no longer processed for sorting by the app. Various small improvements have been made throughout the app.

You can see it here: https://sourceforge.net/projects/ai-file-sorter

AI File Sorter's source code is now up on Guthub. See it here: https://github.com/hyperfield/ai-file-sorter

SoftPedia has added the app to its catalog. You can read the review by Robert Condorache at https://www.softpedia.com/get/File-managers/AI-File-Sorter.shtml

In order to better present the AI Filesorter app to you, this website was created and is now online!

This is the first publicly available version of AI Filesorter, an intelligent tool for ordering your files and directories using AI.