What file formats can TagSpaces preview and how does its AI chat work?

Updated

TagSpaces previews more than 50 file formats directly in the app—images, PDFs, audio, video, spreadsheets, and source code among them—and adds a private AI chat to any folder by connecting to a local model server such as Ollama, llama.cpp, or LM Studio. Both features work on your device without an account or cloud dependency, so the practical conditions are: the file lives in a folder you've opened in TagSpaces, and for AI chat you have a local model server running that TagSpaces can reach.

Previewing files: what to expect

Preview is a built-in viewer, not a full editor. The point is to let you look at a file without launching another application, which matters when you're browsing a tagged collection of mixed content.

The format categories

Category Examples of what you can preview Typical use
Images Common photo and graphic formats Scanning a tagged photo library
Documents PDFs Reading a paper or invoice without a PDF reader
Audio Common audio files Checking a recording before filing it
Video Common video files Confirming which clip a tag belongs to
Spreadsheets Tabular files Glancing at data before opening it in a spreadsheet app
Source code Text-based code files Reviewing a snippet while organizing a project folder

The exact list runs past 50 formats, and the site's own feature page is the place to confirm a specific extension before you rely on it. Two practical notes:

  • Preview is for viewing, not editing. For changes, open the file in its native application.
  • Notes are ordinary files. TagSpaces lets you write notes in Markdown, HTML, or plain text, and each note is a standard file you control—so a note previews like any other text file and stays readable outside TagSpaces.

A concrete scenario

Suppose you're collecting research on a topic. You save web pages with the Web Clipper as HTML, MHTML/PDF, PNG, or URL files, drop a few PDFs and screenshots into the same folder, and tag everything. Later you open the folder and click through the items: the PDFs render in the viewer, the screenshots display, and your Markdown notes open as text. You never leave the app to check what a file is—which is the actual benefit of broad preview support.

How the AI chat works

Each folder can have its own private AI chat. The mechanism is straightforward:

  1. You run a local model server. TagSpaces connects to Ollama, llama.cpp, or LM Studio—software you install and run yourself.
  2. You point the folder's chat at that server. The chat is scoped to the folder, so its context is the files in that folder rather than your whole drive.
  3. You ask for summaries, descriptions, or analysis. The model reads the relevant files and responds.

Because the model runs locally, the processing stays on your machine—the site describes this as fully offline. That's the key difference from cloud AI assistants: no file contents are sent to a third-party service.

What it's good for

  • Summarizing a long document or a set of notes in a folder.
  • Describing files whose contents aren't obvious from the name—useful in a folder of scans or screenshots.
  • Analyzing a group of related files together, such as a batch of research PDFs.

Conditions and limits to plan around

  • You supply the model. TagSpaces doesn't bundle one; the quality and speed of answers depend on the model you run and your hardware.
  • Local models have limits. Very large documents or heavy analysis may exceed what a small local model handles well. Start with a modest folder and a capable model, then scale up.
  • It's per-folder by design. If you want the chat to consider a file, that file needs to be in the folder you're chatting within.

How this fits the rest of the app

Both features reinforce the same design choice: your data stays in plain files on your disk. Tags, descriptions, and colors are stored beside your documents rather than in a database, so they travel with the files across devices and platforms. Preview and AI chat are conveniences layered on top of that file-based structure—if you stopped using TagSpaces tomorrow, your files and notes would still open normally elsewhere.

For syncing across devices, the site points to tools you choose yourself, such as Google Drive, Dropbox, or Syncthing, with encryption options like Cryptomator. TagSpaces is available for Windows, macOS, Linux, and the web, and the site offers a free download and a live demo—the demo is the quickest way to test whether your specific file types preview as you expect before committing to a setup.