The FilmConvert Alternative for Indie Filmmaking: AI Beats LUTs

A FilmConvert alternative for indie filmmaking needs three things Nitrate doesn't offer: shot-matching across a full feature, reference-image grading from a single still, and no $199-per-NLE plugin tax. AI-first grading collapses the first equalization pass on a 1,200-clip festival cut from a weekend into roughly twenty minutes — without touching the creative finish.

I've been a DaVinci Resolve-certified colourist for four years and a cinematographer for longer than that. I've owned FilmConvert Nitrate across Premiere and Resolve since 2019. I've used it on a Puma spot, a handful of music videos, and a short that made the rounds at Slamdance. I still think the Nitrate grain engine is the best in the business. I also think it's no longer the right tool for the way most indie filmmakers actually work in 2026 — and I'll explain why without dunking on it.

Where FilmConvert Nitrate Still Wins

Let me get this out of the way first, because if you skip this section you'll think I'm building a strawman.

The grain. Nitrate's 6K scans of Kodak 5219, Fuji 8553, and the rest of the stock library aren't synthesized — they're photochemical samples digitized at scale. When you put Kodak 5219 grain at 40% on a BMPCC 6K BRAW clip cut against a Sony FX6 clip, both shots inherit the same organic noise structure. That's a real bridge between mixed-camera footage, and it's the closest most of us will get to the texture you felt on Robert Eggers' grade for The Lighthouse — that was actual orthochromatic film stock, but Nitrate gets you in the same emotional neighbourhood.

The price model. $139-$199 per host application, one payment, yours forever. For a colourist who lives inside one NLE and grades one or two projects a year, that math is hard to beat. I won't pretend a subscription is cheaper across five years for that user — it isn't.

So if all you want is film-stock emulation inside Premiere and you're working on a single eight-minute short with cohesive lighting and one camera, Nitrate is still a defensible buy. I'd tell a friend to get it.

The Indie Workflow That Made Me Want Something Else

Here's where it falls apart for the work most of us are actually doing.

The last festival short I cut was 14 minutes, shot on a Canon C70 and a BMPCC 6K with one cutaway day on an iPhone (yes — Tangerine wasn't a one-off, this is normal indie practice now). 412 clips after the assembly cut. Three lighting setups per location, six locations, an interior reception scene with tungsten practicals fighting daylight through a window.

To grade that in Premiere with Nitrate, my workflow was: input transform on each clip manually, build a base grade on the hero shot of each scene, copy the grade to similar shots, hand-tweak every cutaway, then drop Nitrate on top for the film-stock pass. That first equalization pass — just getting all 412 clips into the same exposure and white-balance neighbourhood — took me a full Saturday. Nitrate doesn't help with that pass. It's a finishing layer, not an equalizer.

The other thing: I finish in Resolve. Almost every festival-bound short and feature I know finishes in Resolve, because the festivals want a DCP and the deliverable math is cleaner there. Which means my Nitrate license inside Premiere doesn't follow me to the finish. If I want it in Resolve too, that's another $139. Multiply across a small collective with three editors on different NLEs and you're at $600+ in plugin licenses before anyone has graded a frame.

What AI Shot-Matching Actually Solves

This is the workflow gap I kept hitting, and the reason I started building Leumos AI. Not abstract "AI is the future" reasons — a specific, measurable pain.

When you upload an 80-minute feature edit to a browser-based grader with AI Scene Cut Detection, the timeline auto-chops into shots in roughly two minutes. Match All then equalizes exposure, contrast, white balance, saturation, and hue across every shot — that's the Saturday pass. You don't node-soup it. You don't copy-paste grades down a stack. You drop the clip, hit one button, and your 1,200-clip timeline lands in the same exposure and white-balance neighbourhood automatically.

Then comes the actually creative move: Reference Image Grading. Drop a still from Chayse Irvin's grade on Aftersun — the bathroom-mirror moment, that warm desaturated cyan-magenta separation — and the AI matches your footage to that frame. Not perfectly. Not 100% of the way. But about 70% of the way there, with an intensity slider for taste. The remaining 30% is what a colourist does on a Tuesday. The first 70% is what used to eat a weekend.

That's what Nitrate doesn't do and was never trying to do. Different category of tool.

If you're an indie 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 Honest Limits — What AI Can't Do Yet

I'd lose your trust in a hurry if I pretended this was magic. It isn't.

AI shot-matching struggles with skin tones in mixed light. If your reception scene has a daylight window on one side and a tungsten practical on the other, the model will pick one and over-correct the rest. You'll need Manual Primaries — temperature, tint, exposure — to clean that up the way you would in any grading suite.

Reference-image matching doesn't replicate film-emulation grain. If you want actual Kodak 5219 texture sampled at 6K, that's still FilmConvert's lane. I'm being clear about this because honest scope is more useful to you than overpromising. Match the colour structure of an Aftersun still in Leumos, then layer Nitrate grain on top in Resolve if grain is what you're chasing. They're not mutually exclusive.

AI also doesn't make creative decisions. Day-for-night, brand-Pantone compliance for a sponsor cutaway, the specific desaturation curve that makes a flashback feel like a flashback — those are colourist calls. AI gets you to a clean, equalized starting point in twenty minutes. The taste happens after.

A Side-by-Side for an 80-Minute Festival Feature

Concrete numbers, because comparison pages without numbers are useless.

FilmConvert Nitrate workflow (assuming Resolve finish, ~1,200 clips, mixed BMPCC 6K + FX6): license cost $199 for the Resolve plugin, first equalization pass ~12-16 hours of manual work, no shot-matching automation, film-grain pass ~2 hours, total grade-to-deliverable roughly 5-7 working days for a colourist who knows the tool cold.

Leumos AI workflow (browser-based, same footage): subscription $15/mo Creator or $39/mo Pro, upload + AI Scene Cut Detection ~5 minutes, Match All equalization ~10 minutes, Reference Image Grading for the look pass ~30 minutes, Input Color Space LUT handling the BRAW-to-Rec.709 transform in one click, export to Resolve for finish — roughly 3-4 working days total. No grain emulation built in; bring Nitrate in for that if you want it.

Roughly half the wall-clock time for the actual repetitive work. The creative finishing time is the same — that's not a bottleneck AI should be touching.

When to Use Which (and Why I'm Building Leumos)

If you're shooting an 8-minute one-camera short with one location and you live inside Premiere, buy FilmConvert Nitrate. It's $139 and it will do exactly what you need.

If you're shooting a 60-90 minute festival-targeted feature, or you're a director who also edits and grades because the budget is $20K and there's no one else, or you're cutting four music videos a month and the first equalization pass is eating your nights — that's the indie filmmaker I'm building Leumos AI for. Browser-based, no per-NLE plugin licenses, exports cleanly into Resolve for the finish.

The reason I started building this was simple: I was tired of node soup. I wanted to spend the first hour of a grade looking at the footage and deciding what it should feel like, not setting up the same input transform on 412 clips. AI handles the rote pass. I handle the taste pass. That's the split I want, and that's what Leumos is built around.

Leumos AI launches in mid-2026 — about thirty days from when you're reading this. The first 500 people on the early-access list get 50% off the first year of Creator or Pro. If you're grading a festival short or feature this year and want to skip the Saturday equalization ritual, get on the early-access list here.

Frequently asked questions

Is FilmConvert Nitrate still worth buying in 2026?

Yes, for the right user. If you live inside one NLE, you grade one or two shorts a year, and you want photochemical-quality grain emulation as a finishing layer, Nitrate at $139-$199 one-time is a fair buy. Where it doesn't fit anymore is the indie filmmaker grading a feature with 1,000+ clips across mixed cameras — Nitrate has no shot-matching, no equalization automation, and no reference-image workflow. Buy it for grain texture and film-stock looks. Don't buy it expecting it to handle the first equalization pass on a festival feature, because that's not what it was built to do.

Does Leumos AI replace DaVinci Resolve?

No, and I'd be lying if I said it did. Resolve is still where I finish every festival project — for the node tree, the deliverable management, the DCP export, the audio sync, all of it. What Leumos is built to replace is the first 60-80% of the grade: the equalization pass, the input transform, the look development. You upload BRAW or ProRes RAW, let Match All and Reference Image Grading do the rote work, then export the graded clips back into Resolve for the finishing pass. It's a front-end for your existing Resolve workflow, not a replacement.

Can I use Leumos AI and FilmConvert Nitrate together on the same project?

Absolutely. They solve different problems and stack well. The workflow I'd recommend for an indie feature: upload your edit to Leumos, run AI Scene Cut Detection and Match All for equalization, use Reference Image Grading for the look pass, then export the graded sequence back into Resolve. Once you're in Resolve, drop Nitrate on the timeline as a finishing layer for grain emulation. You get AI shot-matching for the heavy lifting and Nitrate's photochemical grain for the texture. Two tools, one project, no conflict.

What cameras and codecs does Leumos AI support for indie features?

The MVP supports the cameras and codecs that show up on indie sets: BMPCC 6K BRAW, Canon C70 C-Log3, Sony FX6 S-Log3, RED Komodo R3D, and the usual suspects in ProRes RAW. The Input Color Space LUT handles the log-to-Rec.709 transform in one click for each of those gammas. There are upload size caps depending on plan — 400MB on Free, 1GB on Creator, 2GB on Pro — so for a feature you'll be working scene-by-scene or in proxies rather than uploading the entire 80-minute master in one shot. That matches how most indie colourists work anyway.

How is AI shot-matching different from manually building a node tree in Resolve?

A node tree gives you total control and infinite layering — that's why I still finish in Resolve. The cost is time: you build the tree once per scene, copy it to similar shots, then hand-tweak every clip the copy didn't fit. On a 1,200-clip feature, that's where the weekend goes. AI shot-matching skips the build entirely. Match All analyses every clip and pulls them into the same exposure and white-balance neighbourhood in one pass — roughly ten minutes for the whole timeline. You lose granular per-clip control in that first pass, but you get it back the moment you export to Resolve for finishing.

Is a browser-based grader secure enough for an unreleased festival cut?

Fair question — I asked it myself before deciding to build a browser tool. The architecture: uploads are processed in your account, stored encrypted, and auto-deleted on a rolling window (14 days on Creator, 30 days on Pro). You're not publishing your cut to a public URL. The trade-off versus a local plugin is real: a local plugin keeps the file on your drive, a browser tool moves it through a server. For most indie festival work that's an acceptable trade for the workflow speed. If you're under a strict NDA for a studio piece, a local tool is the right call — Leumos isn't trying to be that tool.

Will the export from Leumos match what I see in the browser when I open it in Resolve?

Yes, as long as you handle the colour space transform correctly on the way out. The browser preview is Rec.709, and the export carries the same Rec.709 transform baked in. When you import into Resolve, set your project colour management to match — Rec.709 timeline, Rec.709 output — and the graded clips will look identical to what you approved in the browser. The one gotcha is if your Resolve project is set to DaVinci Wide Gamut or ACEScct and you forget to flag the incoming clips as Rec.709. That's a five-second fix in the clip attributes panel, but worth flagging upfront.

---

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