Software choices in video editing have real consequences: for delivery speed, for how revisions are managed, and for the consistency of the final output. In early 2026 we made the most significant change to our toolset since we started. Here is what we use now, and the reasoning behind each decision.
The current stack
Four tools handle the majority of every project. They were chosen not for their marketing but because they work well together in the specific workflow we’ve built around reliable delivery and clear client communication.
DaVinci Resolve 19, primary editing and colour
Resolve handles all long-form work: YouTube videos, podcasts, documentary-style content, and tutorials. Its node-based colour workflow is the primary reason it replaced our previous NLE for long-form. We can build a colour grade once per channel and apply it consistently across every video, saving significant per-project time while keeping the channel’s visual identity coherent across months of content.
“The best tool is the one your editor knows well enough to disappear inside.”
CapCut Pro (desktop), short-form and Reels
CapCut’s native vertical format tools and caption generation make short-form content significantly faster to produce than adapting a long-form timeline. The caption accuracy is now sufficient that it replaces manual transcription for most projects, requiring a correction pass rather than a full rebuild from scratch.
Frame.io, client review and approval
Frame.io addressed a specific problem: revision communication. Before it, notes arrived via email, voice message, or timestamped documents, all of which required interpretation and carried room for ambiguity. Frame.io lets clients leave frame-accurate, timestamped comments directly on the video. The editor sees exactly what is being referenced. Misunderstandings dropped almost immediately after we switched.
First-draft approval rate across client projects. Frame.io’s timestamped commenting has been the single biggest contributor to reducing revision rounds and shortening the overall delivery cycle.
Descript, transcript-based rough cut
For talk-heavy content, podcasts, interviews, and tutorial-style videos, Descript’s text-based editing is faster than traditional timeline scrubbing for the rough cut stage. Filler removal and structural reordering are done in text, which is substantially quicker. The rough cut then moves into Resolve for the fine cut and colour work.
What changed in 2026 and why
The main change was replacing our previous primary NLE with DaVinci Resolve for long-form work. The decision came down to colour workflow. Our previous tool required building a new grade per video, time that scaled poorly as client volume grew. Resolve’s node structure allows us to build reusable channel-level grades. The setup time per channel is higher initially, typically two or three videos while the grade is refined, but the ongoing per-video time is substantially lower.
- DaVinci Resolve 19, primary NLE and colour grade for long-form
- CapCut Pro (desktop), short-form, Reels, and YouTube Shorts
- Frame.io, client review, annotation, and approval workflow
- Descript, transcript-based rough cut and caption generation
A well-chosen stack compounds over time
The right tools don’t just accelerate individual tasks, they create consistency across a body of work and reduce the cognitive overhead of adapting to each project. That consistency shows up in the output quality and in how reliably a channel’s visual identity holds across months of content.
What we looked at and didn’t adopt
Several AI-assisted editing tools have emerged over the past 18 months. We’ve tested most of them. The consistent pattern: they reduce time on rough cut work for straightforward content, but underperform on the nuanced judgements that make a good edit, pacing, emotional rhythm, knowing when to cut away and when to let a moment breathe. We use AI tools for specific tasks (captions, noise reduction, transcript generation) where they work reliably. We don’t use them to make editing decisions.