Browser-Based Color Grading for Wedding Videography: No Install, No GPU | Leumos AI

Browser-based color grading for wedding videography means uploading multi-cam S-Log3 or BRAW footage from a Sony FX3, Canon R6 II, or BMPCC 6K Pro to a web app — no Resolve install, no dedicated GPU, no node trees. For a 14-hour event with 300+ clips across two or three bodies, the right tool can collapse the first equalization pass from a full weekend to roughly ten minutes per scene.

I've been a colourist for around four years — DaVinci Resolve Certified, BFA in Cinematography, graded ad films for Puma and WHSmith, indie features that hit festivals, and enough music videos to lose count. I've also graded my share of weddings. And I'll say this plainly: wedding work is the most punishing color job in the industry per dollar earned. You're being paid $1,500 to $8,000 to deliver cinema-quality footage on a 4-to-12 week turnaround, on a project where every scene lives in a different lighting universe, shot by two or three operators who didn't quite match their picture profiles.

I started building Leumos because I was tired of node soup. I'd rather create than tweak. This page is for filmmakers who feel the same.

Why wedding footage breaks the old desktop workflow

The reason wedding work eats colourists alive isn't artistic — it's logistical. A typical California wedding day (the kind Henry Robert's reels are built around) gives you somewhere between 250 and 500 clips across two or three bodies. An FX3 on the main angle in S-Log3. A second A7 IV on a gimbal in S-Log3 with a slightly different picture profile because the second shooter forgot to match settings. Maybe a BMPCC 6K Pro on a tripod for the ceremony wide, recording BRAW at constant quality 5.

Now multiply by lighting. The ceremony was outdoors at 3pm under harsh sun. Cocktail hour was golden hour mixed with one warm practical sconce. Reception was a tent with tungsten string lights, magenta LED uplights, and a DJ booth pulsing RGB. Every one of those scenes is a different color problem.

In Resolve you'd build a node tree per scene, copy-paste it across clips in the same scene, then hand-tweak the ones that drift. That's a real day of work. On a 2022 MacBook Pro without a beefy GPU, it's a real day plus a lot of dropped frames during playback. This is the workflow MattWhoTube's tutorials walk through, and it's correct — it's just slow, and it punishes anyone editing on a laptop.

Browser-based grading sidesteps the local-machine problem. The compute happens server-side. Your MacBook Air becomes the color suite.

What "browser-based" actually means in 2026

A few things it doesn't mean. It doesn't mean Resolve running in a browser tab. It doesn't mean a thin client streaming a remote desktop with 200ms latency. It doesn't mean you upload your raw .R3D originals and grade in 4K HDR with full curve nodes.

What it does mean, practically:

• You upload a proxy or compressed master (1080p ProRes Proxy, H.264, DNxHR LB) to a web app
• The AI handles equalization and look-matching at server speed
• You export a graded version, or a CDL/LUT you can apply back in Resolve or Premiere for finishing

For wedding work this is the right level of abstraction. You're not finishing a Netflix series. You're delivering a 4-minute highlight film and a 45-minute documentary edit, both at 1080p or 4K H.265 for the couple to share. The grading decisions are about consistency and a unified look, not pixel-peeping curves at 100% zoom on a $30k reference monitor.

If you're a wedding filmmaker, we're building this for you. Leumos AI launches in ~30 days — join the early-access list and you'll be in the first 500 (50% off the first year).

The three bottlenecks killing wedding turnaround time

After grading a fair number of wedding projects and watching every Daniel J. Holman wedding cinematography breakdown I could find, the same three problems eat every editor's weekend:

1. Multi-cam exposure drift. Two cameras, same wedding, half a stop apart because operator A nudged ISO mid-ceremony. In a 200-shot timeline you're hand-matching exposure dozens of times. None of those decisions are creative — they're janitorial.

2. Mixed-lighting receptions. Golden hour creeping into tungsten and then dipping into LED uplights. A single LUT doesn't survive this. Per-shot white balance does, but at 30 seconds per clip on a long reception sequence, that's 90 minutes you'll never get back.

3. Look cohesion across the whole event. The couple wants the day to "feel like one story." That means the ceremony grade has to talk to the reception grade. The prep montage has to live in the same world as the speeches. A uniformly applied LUT flattens this. Per-shot grading without a unifying tool fragments it.

A realistic browser-based wedding workflow

Here's how I'd run a two-body, 14-hour wedding through a browser tool — the workflow I'm building Leumos around:

1. Out of your NLE, export proxies from your Sony FX3 or Canon R6 II at 1080p ProRes Proxy or H.264, organized by scene (prep, ceremony, golden hour portraits, reception)
2. Upload one scene at a time
3. Apply the Input Color Space LUT to push S-Log3 or BRAW into Rec.709 in one click
4. Let AI Scene Cut Detection chop the upload into a shot-by-shot timeline with thumbnails — no manual node-per-clip setup
5. Pick your reference — a still from a Henry Robert film, a frame from The Wedding Filmer's reception work, or one of your own hero stills — and use Reference Image Grading to push the scene toward it with the intensity slider
6. Hit Match All so every shot in that scene equalizes for exposure, contrast, white balance, saturation, and hue
7. Use Manual Primaries to nudge the handful of shots the AI gets wrong
8. Export graded proxies back into Resolve or Premiere for final delivery and audio finishing

What used to be a node-soup weekend becomes maybe two hours per scene, most of it watching playback rather than building trees.

Where browser-based grading still falls short

I'm building this and I'll still tell you honestly what it can't do. AI color matching struggles with mixed-light skin tones — the bride's face under tungsten warmth and an LED magenta wash at once is a problem no current AI tool handles cleanly. Colourlab, color.io, and fylm.ai all stumble here, and so will any first version of mine. You'll still need to do localized power-window work for those shots in your NLE.

Day-for-night creative grades are also outside the scope. If you want a deeply moody look that fights the source footage rather than equalizing it, AI tools will average toward the reference but won't make the bold creative leaps. That's still a human decision.

And if your couple wants Pantone-exact brand color on a sponsor logo in their reel, you're back in Resolve for that shot.

What I'm building at Leumos

Reference Image Grading, Match All, and AI Scene Cut Detection are the three features I think collapse 80% of the wedding grading workload. The Preset LUT Library with an intensity slider gives you a starting point when you don't have a reference image handy, and the Manual Cut Tool catches anything the AI misses on a tricky cross-dissolve.

Everything runs in your browser. No install. No GPU requirement. Your MacBook Air or a borrowed Chromebook becomes a color suite.

When Leumos launches in ~30 days, pricing will be Free ($0, 2 uploads/day, 400MB max), Creator ($15/mo, 8 uploads/day, 1GB max), and Pro ($39/mo, 20 uploads/day, 2GB max). A wedding videographer running one project a week fits comfortably inside Creator. The first 500 early-access signups get 50% off the first year.

If you'd rather spend Saturday with your family than parked in front of a node tree, join the early-access list.

Frequently asked questions

Do I need a powerful computer to use browser-based color grading for wedding footage?

No. The whole point of a browser-based tool is that the compute happens on the server. A MacBook Air M2, a borrowed Chromebook, or a five-year-old Windows laptop can drive the grading session as long as your internet upload is reasonable. You'll still want a decent screen for color judgment, but you don't need a dedicated GPU, 64GB of RAM, or a desktop tower. This is the main reason I'm building Leumos in the browser — wedding shooters edit on laptops, often on the road between shoots, and the old desktop-color-suite model doesn't match their reality.

Can browser-based tools handle BRAW from my BMPCC 6K Pro?

Not the raw .braw files directly — those are too heavy to upload comfortably from a 256GB SD card and the wrong abstraction for web-based grading anyway. The realistic workflow is to export 1080p or 4K ProRes Proxy or H.264 from your NLE, grade those proxies in the browser, then apply the resulting CDL or graded LUT back to your full-resolution BRAW timeline in Resolve for final delivery. Leumos will support the Input Color Space LUT transform for BRAW directly so your proxies look correct in Rec.709 before the AI starts matching.

How does AI scene matching compare to building a node tree in DaVinci Resolve?

A well-built Resolve node tree is more precise and more flexible — you can pull keys, build power windows, do dimensional grading per shot. AI matching is faster and more consistent across large timelines. For a 4-minute hero film of 60 clips, a great colourist with a node tree wins on craft. For the 45-minute documentary edit of 280 clips, AI matching wins on time and cohesion. Most wedding filmmakers should be using both: AI to handle equalization across the bulk of the timeline, Resolve for the hero shots and any localized fixes.

What about delivering a 4K master to the couple?

You'll still do final export from your NLE. Browser-based grading gives you the look — the graded proxy, a CDL, or an exported LUT — and you apply that look to your full-resolution timeline in Resolve or Premiere for the actual 4K H.265 delivery. This is the same round-trip workflow professional colourists already use with on-set grading and dailies. Browser tools aren't trying to replace your finishing system. They're trying to replace the boring 80% of grading that lives upstream of finishing.

Will my couple's footage be safe on a web app?

Honest answer: it depends on the provider, and you should ask. Look for stated upload encryption (TLS in transit), at-rest encryption, a clear retention window (Leumos will be 14 days on Creator, 30 days on Pro), and a privacy policy that says your footage won't be used for training. I'm building Leumos with all of that — but more importantly, you should treat any third-party tool the way you'd treat a cloud backup: read the terms, know the retention, and don't upload anything you wouldn't put on Google Drive. Most wedding contracts are fine with this; check yours.

Can I still use my own LUT pack from a creator I follow?

Yes. The Preset LUT Library in Leumos will include curated cinema LUTs as a starting point, but you'll also be able to upload your own .cube files — the M.Lookpro packs, your custom builds, whatever you've paid for. Stack a custom LUT under the AI matching layer and you've got the best of both: your signature base look, then AI equalization across the timeline so every shot wears the look consistently. This is how I'd handle a brand-defined wedding shooter who has their own established look.

Is browser-based color grading a Resolve replacement for wedding work?

No, and I'd be lying if I said otherwise. Resolve is still the most capable color tool ever shipped, and for hero shots, complex relighting, or anything requiring power windows and tracking, you should be in Resolve. What browser-based grading replaces is the slow, repetitive equalization pass — the part where you're matching exposure across 200 shots from two cameras. That part has no business being done by hand in 2026. Use the browser tool for the 80% that's tedious. Use Resolve for the 20% that's craft. That's the workflow I'm designing Leumos to fit into.

---

Leumos AI launches mid-2026. The first 500 early-access signups get 50% off the first year. Join the early-access list →