getrapidraw.com

RapidRAW keeps your originals untouched by writing every adjustment to a sidecar file, and it lets you apply those adjustments to many photos at once through batch operations and saved presets. The workflow is built for photographers who want to experiment freely and then repeat a look across a shoot without re-editing each frame. It runs on Windows, macOS, and Linux, and the app itself is about 20 MB.

How the non-destructive workflow works

The core mechanism is simple: RapidRAW never overwrites your RAW file. When you change exposure, tone curves, HSL, detail, or effects, those settings are saved separately in a sidecar file that sits alongside the original. The image you see on screen is a rendering of the original plus the sidecar instructions.

That has three practical consequences:

• Your originals stay intact. You can always return to the untouched file, whether you want to start over or hand the photo to another tool.
• Edits are reversible and portable. Because the changes live in a sidecar rather than baked pixels, you can revisit an edit later and adjust any slider without quality loss.
• Deleting or moving files needs care. Since the sidecar is a separate file, keep it with its original when you reorganize folders. If you move a RAW file without its sidecar, the edit won't follow.

All adjustments are processed on the GPU, so the preview updates in real time as you drag sliders. That matters for a non-destructive workflow: you're evaluating the rendered result constantly, and slow feedback makes experimentation painful.

A concrete example

Say you shoot 200 photos at an event and want a consistent warm, slightly contrasty look. You edit one frame — adjust white balance, pull a tone curve, tune the HSL mixer, add a touch of detail enhancement — and save those settings as a preset. Your original files are unchanged; only sidecar data records the look. You then apply that preset across the rest of the set in a batch, and each photo gets its own sidecar. If one image looks off, you open it and tweak only that frame without affecting the others.

Managing a collection in the image library

RapidRAW includes a complete image library so you can browse and manage your whole photo collection in one interface rather than jumping between folders in a file manager. Combined with the sidecar model, this means your edits travel with your library view: you can scan a shoot, spot the frames that need work, and apply changes without leaving the app.

For large collections, the library is where batch work starts — you select the images you want to treat the same way, then push adjustments or a preset to all of them.

Batch operations and presets

Batch operations apply adjustments to multiple images simultaneously, and the preset system is what makes batching practical: you save a favorite look once and apply it instantly to any photo.

A sensible order of operations: cull in the library first, edit a representative frame, save the preset, batch-apply, then review and correct individual images. Doing it in that order avoids re-batching after you've made per-image fixes.

Where AI fits into the workflow

AI masking (subject, sky, and foreground detection) and automatic image tagging are built in and run locally on your computer — fast, free, and with no setup. That means localized edits, such as brightening a sky or lifting a subject, can be part of your batch workflow without sending files anywhere.

Generative tools like object removal are different: they need an optional ComfyUI backend (self-hosted on your own GPU) or RapidRAW Cloud, an optional subscription service. If your batch workflow depends on generative edits, plan for one of those two paths; the built-in AI tools alone cover masking and tagging, not generative removal.

Common snags to plan for

• Sidecar discipline. Back up sidecars together with originals, and don't separate them when moving files between drives.
• GPU dependence. Real-time feedback relies on GPU acceleration, so performance on weaker hardware will differ; test with a representative batch before committing to a large job.
• Generative edits aren't local by default. Built-in AI is local, but object removal requires ComfyUI or the Cloud subscription — decide which before building a workflow around it.
• Presets are starting points. A preset applied across mixed lighting will need per-image correction; batch first, then refine.

If you want to go deeper, RapidRAW publishes documentation covering setup, masks and AI, presets, and complete edit walkthroughs, plus a shortcuts reference for speeding up the repetitive parts of batch work.

RapidRAW offers three ways to handle AI editing: built-in AI tools that run locally with no setup, a self-hosted ComfyUI connection for advanced generative edits on your own hardware, and an optional Cloud subscription for the same generative results without the setup. Choose built-in AI for masking, tagging, and simple object removal; choose ComfyUI if you have a capable GPU and want full control over models and workflows; choose Cloud if you want generative edits without managing a server and are willing to pay a subscription that funds development.

The Three Options at a Glance

Built-in AI: The Default for Everyday Edits

The built-in tools are integrated directly into RapidRAW and run entirely on your machine. According to the site, they are fast, free, and require no setup. They cover three jobs:

• AI masking with subject, sky, and foreground detection, which you can combine with other masks for localized edits.
• Automatic image tagging to help organize a library.
• Simple object removal for basic cleanups.

This is the right choice if your AI needs are masking and light retouching, or if you simply don't want to maintain any extra infrastructure. The trade-off is scope: these are fixed tools rather than a general generative pipeline.

Self-Hosted ComfyUI: Maximum Control for Capable Hardware

If you have a capable GPU and want more than the built-in tools offer, you can connect RapidRAW to your own local ComfyUI server. The site describes this path as using your own hardware and models, with cost-free advanced generative edits and custom workflow selection.

Choose this option when you already run ComfyUI, want to pick specific models, or need generative edits beyond simple object removal. The costs are practical rather than monetary: you need suitable hardware, and you take on the setup and maintenance of the server and its connection to RapidRAW.

Cloud Service: Convenience With a Subscription

The Cloud option is for people who want the same high-quality generative results as self-hosting without any hassle. It is an optional subscription, and the site notes it directly funds RapidRAW's development. The stated benefits are maximum convenience, no setup, the same results as self-hosting, and support for the project.

Choose Cloud if you don't have a capable GPU, don't want to run a server, or would rather pay to support development than manage infrastructure. The trade-off is the recurring subscription cost and less control over the underlying models compared with self-hosting.

How to Decide

• You mainly need masks, tagging, and basic object removal → use the built-in AI tools. No setup, no cost.
• You have a capable GPU, already use ComfyUI, or need specific models and workflows → self-host. You get the most control at no edit cost, in exchange for setup and hardware.
• You want generative edits without any setup and are comfortable with a subscription → use Cloud. You get self-hosting-level results with none of the maintenance, and you support the project.

The three options are not mutually exclusive in practice: the built-in tools remain available regardless, and the generative choice comes down to whether you'd rather supply hardware and configuration or pay for a managed service.

RapidRAW's official site (getrapidraw.com) is the starting point for all three: the Docs section hosts the full guide from first setup through advanced masking, presets, and workflows; a dedicated Shortcuts page lists the essential hotkeys; and a Tutorials section provides complete photo edits with before/after walk-throughs. All are reachable from the site's "Master your editing workflow" area, where the Docs, Shortcuts, and Tutorials entries sit together.

What each resource covers

The site describes the guide as built to be focused rather than exhaustive — it moves from initial configuration straight into masking, presets, and workflow topics, so you can read it linearly or jump to the section you need.

Suggested path through the material

1. Start with Docs — first setup. Confirm your install and get oriented in the interface before touching edits.
2. Read the Shortcuts page early. The hotkeys are described as what makes RapidRAW "incredibly fast," so learning them before you build habits saves rework later.
3. Work through a Tutorial. Follow one complete edit with its before/after so you see how adjustments, masks, and presets combine on a real image.
4. Return to Docs for advanced masking and presets. This is where the guide covers combining AI subjects, skies, and Flow Masks for localized edits — the part most worth reading after you've done a full edit once.

The masking lesson worth prioritizing

The Shortcuts section's companion topic — combining AI subjects, skies, and Flow Masks — is the highest-leverage thing in the documentation. RapidRAW's AI masking detects subjects, skies, and foregrounds automatically, and the docs explain how to combine those masks with Flow Masks for precise local adjustments. If you only read one advanced section, make it this one: it's where automatic detection and manual control meet, and it's the workflow most likely to change how fast you edit.

If you want more than the official material

The site notes RapidRAW has been featured or discussed by Gizmodo, The Phoblographer, CHIP, YouTube channels, Hacker News, and Reddit. Those third-party write-ups and videos are worth checking for independent walk-throughs and real-world impressions, but treat the official Docs as the authoritative source for how features are meant to work — third-party coverage may lag behind releases.

Practical notes

• The app is a roughly 20 MB download, available via the site's Download button and on GitHub, for Windows, macOS, and Linux.
• Edits are non-destructive, stored in a sidecar file, so you can follow tutorials on your own photos without risking originals.
• Generative AI tools (such as object removal) depend on your chosen backend — built-in, self-hosted ComfyUI, or the optional Cloud service — so a tutorial that uses generative features may assume a setup you haven't configured yet. Check which AI path a tutorial relies on before following it step by step.

RapidRAW runs on Windows, macOS, and Linux, and its editing engine is GPU-accelerated, so a reasonably capable GPU is the main hardware consideration. The built-in AI tools (subject/sky/foreground masking, automatic tagging, simple object removal) run locally on your own computer with no setup, while the optional ComfyUI route is aimed at users with a capable GPU who want to run their own local server and models.

Supported Operating Systems

The download is listed at just 20 MB, which the site notes is lighter than a typical RAW image. The same build family is offered across all three platforms; there is no indication of a platform-specific feature split.

What Actually Matters: The GPU

Because all image adjustments are processed on the GPU for real-time feedback, the GPU is the component that determines how smooth editing feels. The site frames GPU acceleration as central to the product rather than an optional extra, so plan around a machine with a discrete or otherwise capable GPU if you intend to edit large RAW files or use AI masking heavily.

The site does not publish minimum GPU models, VRAM figures, or CPU/RAM baselines. If you need exact thresholds before committing, check the documentation or the project's GitHub page rather than assuming a number.

The Three AI Setups and Their Requirements

RapidRAW lets you choose how AI features run, and the requirements differ by choice:

Built-in AI tools

• Runs entirely on your computer.
• Described as fast, free, and requiring no setup.
• Covers AI masking (subject, sky, foreground), automatic image tagging, and simple object removal.
• Best fit: most users who want AI features without configuring anything.

Self-hosted via ComfyUI

• Requires a capable GPU and your own local ComfyUI server.
• Lets you use your own hardware and models with a custom workflow.
• Described as cost-free advanced generative edits, but the setup and compute are on you.
• Best fit: users with strong local hardware who want maximum control over generative edits.

RapidRAW Cloud

• An optional subscription service for the same high-quality results with no setup.
• The site states it produces the same results as self-hosting and directly funds development.
• Best fit: users who want generative tools without local GPU overhead or configuration.

Note that the built-in tools are described as free, while Cloud is explicitly a paid optional subscription. No prices are listed on the page, so treat cost as something to confirm at signup.

Other Practical Requirements

• Storage for originals plus sidecars: edits are non-destructive and stored in a sidecar file, so your original images stay untouched. Budget disk space for both the photos and the sidecar files.
• A photo library location: the built-in image library manages your whole collection in one interface, so decide where your catalog lives before importing.
• Optional ComfyUI install: only needed if you pick the self-hosted AI path.

Quick Decision Guide

• Just want to edit RAW files on any of the three OSes: download the ~20 MB build and use the built-in GPU-accelerated tools.
• Have a strong GPU and want custom generative workflows: connect RapidRAW to your local ComfyUI server.
• Want generative edits without hardware or setup: use the optional Cloud subscription.

Common Stumbling Points

• Assuming the app is lightweight in every sense: the download is small, but GPU-accelerated editing still depends on your graphics hardware.
• Expecting published minimum specs: the site doesn't provide them, so verify against the docs or GitHub before buying hardware for it.
• Confusing the AI tiers: built-in tools are local and setup-free; ComfyUI is local but requires configuration; Cloud is remote and subscription-based.

RapidRAW is a GPU-accelerated, non-destructive RAW image editor for Windows, macOS, and Linux, positioned as a lightweight, modern alternative to Lightroom. It suits photographers who want fast, real-time editing feedback, AI-assisted masking, and a built-in image library in a small download (about 20 MB) — and it is worth considering if you also want control over where AI processing runs: locally, on your own ComfyUI server, or through an optional cloud subscription.

Core editing model

RapidRAW is built around three properties that define how you work with it:

• GPU-accelerated processing. All image adjustments are processed on the GPU, so edits render with instant, real-time feedback rather than waiting on CPU-bound previews.
• Non-destructive workflow. Every edit is stored in a sidecar file, leaving your original RAW files completely untouched. You can revisit or discard adjustments without risk to the source image.
• Cross-platform availability. The same editor runs on Windows, macOS, and Linux, which matters if your workflow spans more than one operating system.

Key capabilities

The AI masking tools are the most workflow-defining feature: automatic subject, sky, and foreground detection removes much of the manual brush work behind localized edits, and combining masks lets you build layered selections (for example, a sky mask refined by a foreground mask) without leaving the editor.

Three ways to run AI features

RapidRAW's AI is deliberately flexible, and the choice affects cost, setup effort, and privacy:

Built-in AI tools

Integrated directly into RapidRAW and running entirely on your computer. These are described as fast, free, and requiring no setup. They cover AI masking (subject, sky, foreground), automatic image tagging, and simple object removal.

Self-hosted via ComfyUI

For users with a capable GPU who want maximum control, RapidRAW can connect to your own local ComfyUI server. This lets you use your own hardware and models, get advanced generative edits at no cost, and select custom workflows. The trade-off is setup and hardware requirements.

Cloud service

An optional subscription provides the same high-quality results with no setup, and directly funds RapidRAW's development. This is the simplest path if you don't want to manage local AI infrastructure.

The practical decision: use built-in tools for masking and tagging, self-host if you already run ComfyUI and want full model control, and choose the cloud option if convenience matters more than running AI locally.

Where it fits — and where to check first

RapidRAW is a strong fit if you want a fast, small-footprint RAW editor with real-time GPU feedback, non-destructive sidecar-based editing, and AI masking without a heavyweight install. Its batch operations and preset system also make it reasonable for volume work such as event or travel shoots.

Before committing, verify the specifics that matter to your workflow: the exact list of supported RAW camera formats, the full adjustment toolset beyond the headline features, and the current terms of the cloud subscription. The site's documentation covers setup, shortcuts, masks and AI, and complete tutorial walk-throughs — a sensible place to confirm fit before migrating an existing catalog.

If your priority is a mature ecosystem of plugins and cloud sync, a traditional desktop editor may still be the safer default. If your priority is speed, a tiny install, cross-platform support, and control over where AI runs, RapidRAW is designed precisely for that.