A walkthrough of FilesDesk's free tool suite. All client-side, all useful for the kind of small file jobs that come up mid-workflow.
There's a specific category of task that always shows up at the worst time. You're three commits deep into something, and suddenly you need to strip GPS data from a screenshot, or verify a download's SHA-256, or rename 60 files that came in with garbage names. The flow break is annoying. The bigger annoyance is the dance of finding a site that does the job without making you upload the file, sign up, dismiss a cookie banner the size of a small novel, and dodge three popups.
I came across FilesDesk's free tools page recently and ended up bookmarking the whole thing. It's a set of eight browser tools that run entirely client-side. No upload, no signup, no server doing the work. Just JavaScript in the tab.
URL is filesdesk.app/tools if you want to follow along.
I'll go through each one with the honest version of when I'd actually reach for it, what's interesting about the implementation, and where it falls down.
1. EXIF Viewer and Remover
Reads metadata out of JPG, PNG, WebP and HEIC files. Shows you everything in there: camera, lens, exposure, GPS coordinates, timestamps, copyright fields. One button to strip it all.
The dev-relevant use case is screenshots and product photos. If you're shipping documentation or a blog post and you grabbed a photo with your phone for a quick illustration, that photo carries the GPS of wherever you took it. Stripping EXIF before publishing is a habit worth building. Same goes for any image you're attaching to a public bug report or pasting into a Slack channel you don't fully trust.
For the inspection side, it's also a useful debugging tool when you're testing image upload pipelines. You can quickly verify what metadata is making it through your processing chain by checking before and after.
2. Bulk File Renamer
Up to 200 files at a time. Find and replace, sequential numbering, prefix and suffix, regex. Outputs as a ZIP.
I know what some of you are thinking. I can do this in five lines of bash. Yes, and so can I, but the value of this tool is not for the days when I'm at my own machine with my own scripts. It's for the days when I'm helping a non-technical colleague, or working from a borrowed laptop, or doing the rename inside a sandboxed environment where I don't want to install anything.
The regex support is the part that elevates it past the basic web rename tools. You can do real capture group rewrites, which covers most of what people actually script for this.
The 200 file cap is a memory call, not arbitrary. Holding more than that in a browser tab and zipping it gets unhappy. Fine for the kind of batch sizes humans actually face.
3. PDF Metadata Viewer and Remover
PDFs carry their own metadata. Title, Author, Subject, Keywords, Creator, Producer. Word and Pages both quietly drop your full name into Author. Browsers sometimes embed the source URL. Some PDF generators include the software version and build, which is mild information disclosure if you're handing the file to someone you'd rather not give a tech profile to.
This tool lets you read those fields and either edit them or wipe them. For anyone writing about security or sending PDFs to clients, it's a small habit that closes a real gap. Five seconds of work, no detectable downside.
4. HEIC to JPG Converter
HEIC is technically a better format than JPG. It's also a constant source of friction the moment a file leaves the Apple ecosystem. Windows still doesn't ship with HEIC support out of the box. CI pipelines and image processing libraries written years ago often don't handle it. Forums and form uploads silently reject it.
Bulk convert in the browser, quality slider on the output. The interesting bit technically is that HEIC decoding in JavaScript is non-trivial because the format uses HEVC, so doing this client-side at reasonable speed is more impressive than it sounds.
I use this almost exclusively when someone on an iPhone sends me a batch of photos for some project and I need them in a format my tooling actually understands.
5. File Hash Generator
MD5, SHA-1, SHA-256, SHA-512 for any file you drop in. Computed using the Web Crypto API where supported.
Use cases that come up for me:
- Verifying that an ISO or installer matches the publisher's checksum before I run it
- Confirming two files are byte-identical when they came from different sources
- Spot-checking that a build artifact hasn't changed when I think it shouldn't have
- De-duplicating downloads that have different filenames
The thing I like about doing this in a browser is that it works on any OS without me needing to remember the platform-specific incantation. On macOS it's shasum -a 256. On Linux it's usually sha256sum. On Windows it's Get-FileHash in PowerShell or certutil if you're feeling old school. Drag the file into a tab and read the result.
6. Image Compressor
JPG, PNG and WebP with a quality slider. Client-side re-encoding, no upload.
The relevant scenarios are pretty much any time an image has to travel somewhere and size matters. Screenshots before pasting into a chat tool that doesn't auto-compress. Photos before attaching to a Notion page that's already heavy. Images for a blog post where Lighthouse is going to scold you about transfer size.
The quality slider matters because the right setting really does depend on the source. A screenshot of mostly flat colour will tolerate aggressive compression. A photograph with gradients and noise won't. Giving you the dial rather than picking for you is the right call.
7. Filename Sanitizer
This is the most niche tool and the one I'd defend most strongly. Different operating systems have different rules about what's allowed in a filename. Windows forbids < > : " / \ | ? * and reserved names like CON and PRN. macOS is mostly permissive but has its own quirks. Cloud sync services have stricter rules than either OS. Emoji in filenames will break Git on some systems and not on others.
If you've ever tried to clone a repo and seen "invalid path" errors, or watched a Dropbox sync just refuse to copy certain files with no explanation, this is the underlying issue.
The sanitizer strips problem characters in batch. You set the rules (replace spaces with underscores, drop emoji, normalise punctuation) and get a clean set of filenames out. Cross-platform safe.
I'd argue this should be a Git pre-commit hook for some teams. Until that exists, the browser tool covers it.
8. OCR: Image to Text
This is the standout. Tesseract.js running in a Web Worker, thirteen languages supported. Drop in an image, get text out.
Tesseract is the open source OCR engine that's been around forever, originally from HP and now mostly maintained by Google. Getting it running well in a browser via WASM is a genuine engineering feat. Accuracy is comparable to what you'd get from cloud OCR services for clean source material.
Realistic uses I've hit recently:
- Extracting an error message from a screenshot a teammate sent in Slack so I can search for it
- Pulling text out of a scanned PDF that turned out to be image-only
- Copying configuration values from a photo of a sticker on a piece of hardware
- Reading a chunk of code from a screenshot when the original gist link is dead
For terminal screenshots in particular it's nearly perfect, which is more useful than you'd expect once you start noticing how much code lives in image form online.
Why I'm into this set specifically
A few things make this collection worth bookmarking rather than just being another grab bag of utilities.
The privacy story is real, not marketing. There's no server endpoint receiving the file. You can verify this in DevTools by watching the Network tab while you use any of the tools. The request count stays flat. That's a different category from sites that claim privacy but then upload your file to "process it securely".
Loading speed is good because the pages are doing real work in the browser, not waiting on round trips. Once the page is loaded, the tool runs at the speed of your local CPU.
No tracking-heavy junk in the way. No popup asking for your email before you can use the tool. No watermark on outputs. No artificial size limits beyond what a browser tab can actually handle in memory.
The honest tradeoff: this is the right pattern for small to medium jobs done occasionally. If you need to process ten thousand files in a nightly cron, write a script. If you need to OCR an archive of a million PDFs, use a real pipeline. For the one-off and small-batch tasks that show up in real work, the browser tools are the path of least resistance.
FilesDesk itself is a paid desktop app for AI-powered file renaming, which is the obvious next step up from the free tools. The team has been adding to the free set steadily, and they take suggestions at the contact page. If there's a small file task you keep doing manually and it's not on the list, that's a reasonable bug report to send.
Bookmark, don't memorise
The reason these are worth saving rather than just remembering is that you'll forget they exist exactly when you need them. The next time you're about to upload a private photo to some random EXIF site or about to install ImageMagick on a fresh laptop just to compress one screenshot, that's the moment to reach for the bookmark.
Eight tools, all free, all running locally, none of them trying to extract a signup from you. That's a rarer combination on the open web than it should be.