About Bitgrain

Bitgrain is a browser-based creative suite built without any AI in it. It is a photo lab, a print shop, a layered design studio, a brand toolkit and a video editor, and every effect in it is a classical image-processing algorithm rather than a model. It is made and maintained by Diptanshu Mahish, working independently from India.

Why it exists

Around 2024 nearly every design tool started shipping the same feature: a text box that generates the work for you. The pitch was that describing an outcome is faster than making it. That is true, and it is also the part of the job most people got into design to do.

Bitgrain was started from the opposite premise. The tools here are levers, not prompts. You choose a dither algorithm, set a screen frequency, pick an ink, decide how far to crush the midtones. The software does the arithmetic and nothing else. If the result is good, it is because of a decision you made, and you can explain every one of them.

There is a practical argument as well as an aesthetic one. Deterministic algorithms are reproducible. The same image with the same settings gives the same pixels today, next month and in five years. Nothing regenerates differently because a model was updated. For anyone producing a series, a catalogue or a set of covers that have to look related, that consistency is the whole point.

What it does

The photo side develops around thirty camera RAW formats in the browser, along with HEIC and HEIF from iPhone, TIFF and DNG. There are thirty eight film and cinematic LUTs, genuine HDR still export in BT.2020 and PQ, and batch processing that runs a whole shoot through one effect chain.

The print side is built for real production rather than screen previews: risograph ink separations exported as per-ink layers, true CMYK four-channel halftone separations, newsprint and zine finishes, screen-print flats and pen-plotter SVG with real vector paths.

The studio is a multi-page layered document editor with custom typography, a vector pen, design tokens and templates, exporting to PNG, WebP, SVG and PDF. The video timeline handles trim, split, speed, beat detection, transitions and per-clip audio EQ, rendering deterministic MP4 up to 4K.

Across all of it sit the effects: fourteen dither algorithms, halftone in mono, colour and CMYK, ASCII and Braille conversion, Sobel edge detection, crosshatch, ink wash, manga screentone, glitch, duotone, grain and twenty four hardware and print palettes.

How it is built

Everything runs in your browser using WebAssembly and Canvas. Image processing happens on your machine, so photographs are not uploaded to a server to be developed, screened or exported. That is a design decision rather than a marketing line: the processing pipeline genuinely has no server step in it.

Projects save as a portable .btg file that carries layers, effect settings, type and palette together, so a project can be handed to somebody else or reopened much later and still produce the same output.

The effects engine is written by hand. There are no models, no training data, no inference and no prompts anywhere in the product, and there is no plan to add any.

Who makes it

Bitgrain is built by Diptanshu Mahish, an independent developer and designer based in India. It is not venture funded and there is no team behind it beyond one person, which is why the product moves in the directions it does and why support replies come from the person who wrote the code.

It is free to start. The free tier exports at web resolution with a watermark and has no cap on how many exports you make. Pro is a one-time payment covering ten years, which removes the watermark and unlocks full resolution export, SVG vector, video export up to 4K, batch export and cloud sync. It is not a subscription, there is no auto-renewal and there is nothing to cancel.

If you want to reach the person who makes it, the contact page has the direct routes.

Common questions

Does Bitgrain use AI anywhere?

No. Every effect is a classical, deterministic image-processing algorithm written by hand. There are no neural networks, no generative models, no inference and no prompts in the product, and none are planned.

Are my images uploaded?

No. Image processing runs in your browser with WebAssembly and Canvas, so photographs are developed, screened and exported on your own machine.

Who runs Bitgrain?

Diptanshu Mahish, working independently from India. It is not venture funded and there is no wider team.

Is it free?

Yes to start. The free tier exports at web resolution with a watermark and no export limit. Pro is a single payment covering ten years, not a subscription.

Elsewhere