Aihangsoft is an independent project that builds 44 file tools for images, PDFs, video and audio. Forty of them run entirely inside your browser, so your file is never uploaded. The other four are Cloud AI tools that do upload, and they are the only part of the site that costs money.
There is no company behind it, no investors and no data business. The browser tools are free and unlimited because they cost nothing to run: the work happens on your own processor, inside a tab, with no upload queue and no server in the middle. That one design decision explains most of the choices on this page, including what the tools can and cannot do. It also sets the honest limits described further down, because a browser is a real constraint rather than a marketing line.
What Aihangsoft actually is
Aihangsoft is a set of small, single purpose tools rather than one large editor. Each page does exactly one job, opens instantly, and asks for nothing before you start. There is no workspace to learn and no project to set up.
The 44 tools are grouped into five families:
- Image tools (7). Compress, convert, crop, watermark, upscale, read text from an image, and strip EXIF metadata.
- PDF tools (13). Merge, split, build a PDF from images, convert to JPG, reorder pages, encrypt, decrypt, lossless optimisation, extract text, add page numbers, add a watermark, sign, and edit metadata.
- Video and recording (13). Compress, convert, trim, merge, rotate, change speed, watermark, make a GIF, extract frames, burn in subtitles, convert subtitle formats, record the screen, and record the webcam.
- Audio tools (7). Extract audio from video, convert format, trim, merge, compress, raise the volume, and record a voice memo.
- Cloud AI tools (4). Remove a background, erase an object, enhance a photo, and separate vocals from a track.
Why the files stay on your device
Most online converters upload your file to a server, process it there, and hand back a download link. That model creates a copy of your document on someone else's disk, which then has to be deleted on a schedule. It also means the site can see your file, and it needs a per day limit to control its own costs.
The 40 browser tools avoid all of that by doing the work where the file already is. When you drop a PDF into the compressor, the bytes are read into the page, processed by a WebAssembly module running on your machine, and written back out as a download. Nothing is sent anywhere, so there is no retention window to explain and no daily counter to hit.
The four Cloud AI tools are the exception, and they say so on their own pages before you upload anything. Background removal, object erasing and photo enhancement need models far larger than a browser tab can hold, and vocal separation runs a heavier model still. Those runs cost real money, which is why only they consume credits. Any file you send to them is processed by Volcengine AI MediaKit and deleted on a timer.
What is running inside the browser tab
Running this much locally is only possible because the browser has become a serious runtime. Each tool family uses a different engine, and all of them are open source libraries that the page loads on demand rather than on every visit.
- Images. The Canvas API handles most conversions and resizing. Lossless PNG re-compression uses OxiPNG compiled to WebAssembly.
- Image upscaling. UpscalerJS with TensorFlow.js runs an ESRGAN super resolution model locally. The model files are downloaded once and then cached by the browser.
- PDF. pdf-lib handles authoring tasks such as merging, page numbers and signatures, while PDF.js renders pages for previews and text extraction. Encryption, decryption and lossless optimisation use qpdf compiled to WebAssembly.
- Video and audio. Every video and audio tool is powered by ffmpeg.wasm, a WebAssembly build of ffmpeg. It is the same encoder that desktop software uses, running in a tab.
- Text recognition. OCR uses tesseract.js. The English language data is served from this site so the first scan does not depend on a third party.
- Recording. Screen and webcam capture use the browser's own MediaRecorder, getDisplayMedia and getUserMedia interfaces, which is also why you have to grant permission each time.
Because those engines are large, the video and audio tools download the ffmpeg core the first time you open one. That first run is noticeably slower than later ones, and the browser keeps it cached afterwards.
What is free, and what costs credits
All 40 browser tools are free forever, with no account and no usage limit. They cost nothing to operate because they consume none of our resources, so there is no reason to meter them. You do not need to sign in, and no watermark is added to anything you make.
The Cloud AI tools are metered, because each run is paid for. A free account receives 5 credits a day. Pro is $6 a month, or $45 a year, and includes 1,500 credits a month. You can also buy a pack of 500 credits for $9 that never expires. A background removal, an object erasing run and a standard enhancement each cost 1 credit, a professional enhancement costs 6, and an advanced enhancement costs 40.
The exact plan details, what each credit buys, and how the free allowance resets are all on the pricing page.
What Aihangsoft deliberately does not do
A tool that claims to do everything is usually hiding something. These are the limits that matter, stated plainly so you can check them against your own needs before you start.
- No HEIC decoding in the browser tools. Browsers cannot decode HEIC, the format used by recent iPhones, so the local image tools cannot open those files at all. Cloud AI pages decode HEIC on the server instead.
- No video resizing or resolution scaling. There is no tool that changes a video's width and height. Trimming, compressing, rotating and converting are supported, rescaling is not.
- No cloud storage for the browser tools. Because nothing is uploaded, there is also nothing to save to an account. Files exist only in the open tab and disappear when you close it.
- No editing suite. There are no layers, no timeline and no multi-track mixing. Audio merging joins clips end to end rather than blending them.
- No fixed output size target. The image and PDF compressors lower a quality or efficiency setting rather than promising an exact result such as 100 KB. Claiming otherwise would be dishonest.
Who runs Aihangsoft
Aihangsoft is an independent project operated by an individual developer rather than a registered company. There is no team to contact and no sales department, just one person who built the tools and maintains them.
For a bug report, a feature request, a billing question or a privacy request, email [email protected]. Questions about data handling are answered in more detail in the privacy policy.