Supacut

Supacut's core promise is narrower than most AI editing tools claim, and that's the point. It does not cut your film. It organizes the material you would otherwise log by hand. What it actually does well: import keeps your original camera files in place, per-interviewee, so nothing is moved or transcoded into a proprietary store. Analysis produces thematic pillars — "the main areas of material found across your interviews," as the interface itself labels them — with each clip carrying its speaker, timecode range, and original quote. That means you're not reading an AI summary and guessing at the source; you click Play on "Maya Chen 12:04–12:31" and you're looking at the footage. The compare step across interviews is where the tool earns its keep on a documentary with fifteen subjects talking around the same event, and the built-in review player lets you audition a shortlist in sequence with reordering before any of it lands on a timeline. Re-running the analysis as your understanding of the footage shifts is a real workflow, not a one-shot export. The bring-your-own-key architecture is the most consequential design decision here. The AI runs through your own model-provider account, so you pay the provider directly instead of buying credits or hitting a usage cap from Supacut, and the app itself is $39 per year billed annually with updates included. For a solo editor or a small shop this keeps the cost floor genuinely low and removes the meter-anxiety that comes with credit-based AI tools. The tradeoff is that quality depends on the key and model you choose, and you own the setup. The honest limits: it does not transcribe, so you need existing transcripts before Supacut is useful at all. It exports XML with in/out points to Adobe Premiere Pro, DaVinci Resolve, or Final Cut Pro — there's no Avid Media Composer path, which rules it out for shops standardized on Avid. There's no real-time collaborative editing or shared project seats, so it's a single-editor tool in practice. And the pillar suggestions are deliberately framed as areas of material rather than a story structure: if you want software to hand you a narrative, this isn't it, and that's a feature of the design rather than a gap. Where it fits: documentary editors sitting on hours of interviews, branded and corporate producers turning talking heads into a narrative, story producers who want pillars and soundbites organized before the timeline. Where it doesn't: scripted fiction with no interview material, anyone whose primary need is transcription or captioning, and Avid-based teams.