Narrative

Narrative is an AI-first video editor built around a simple idea: bring your footage, say what you want made, and cut it together with an editor that does the work with you. Rather than assembling cuts by hand, you upload clips, describe the edit you want in plain words, and keep refining the result through chat. Narrative brings video editing, custom motion graphics and reference-video style matching into one interface. It builds a finished draft on a timeline you can still open and change, so the output is not a locked black box but a real project you can adjust. It is designed for people who already have footage and a clear idea of the outcome, but who do not want to learn Premiere or After Effects to get there. From podcast clips to launch videos, the first step is the same: say what you want. The problem Narrative addresses is the gap between having footage and having a finished cut. Editing traditionally means learning complex software, understanding timelines and motion graphics, and spending hours scrubbing through material to find the good bits. One producer notes that every episode used to cost an afternoon of finding the good bits. A media lead describes the difficulty of finding every goal in a ninety minute match, including the one the camera nearly missed. A product marketing team says their unboxing demos went from a freelancer and a week to a sentence and ten minutes. Narrative targets exactly that gap: the time, cost and specialised skill that sit between raw footage and a usable edit, whether the material is a two hour shoot, a wedding, a podcast or a match recording. The core of the product is editing by talking. You ask in ordinary language, and the agent does the work. An example from the site is a request to make a 15 second reel of the best rides, with a title at the start and music under it. The agent then reads the transcript, edits the clips and reports back what it did, explaining that four rides were placed on the strip with the tightest one first, the title running from 0:00 to 0:03 and the music bed sitting under everything. Anything you would say to an editor, you can say to this one — it works from the words, not a menu. The site lists prompts such as cutting to the beat, making it feel like a trailer, putting every goal in order, removing the ums, going vertical with captions, starting on the best line, tightening to thirty seconds, warming it up a little, and putting the title back at the end. Everything lives in the same editor as the conversation: the transcript, captions, versions, the frame and the models behind it. The editing surface is the editor itself, not a picture of it — you can press play, scrub the strip and mute a track. Tracks and elements shown in the interface include graphics, clips, dialogue and music, along with elements such as a title and a location card, with individual clips listed by timecode and duration. You can add music, sound effects and transitions, and Narrative also supports custom motion graphics and reference-video style matching, so you can take inspiration from reference videos. Because the assistant reads the transcript while cutting, requests such as removing filler words or placing captions are grounded in what was actually said in the footage. Every turn is a version. The version history in the interface shows a numbered list of states, with the current one marked and older entries available to restore — for example v12 as the current turn, v11 with edited styles, v10 where a clip was trimmed, and v9 where the frame was set to 9:16. Each entry carries a restore action, so you can step back to any of them or put the project back to how a turn left it. A completed turn also reports what it used, such as the number of tools and credits consumed. For teams, this matters: one testimonial notes that with every version kept, nobody on the staff can break the cut. You can choose the speed, or the brains. Narrative offers three model tiers: Fast, for quick cuts and small changes; Balanced, for most edits most of the time; and Max, for long footage and hard briefs. The frame control handles aspect ratio, starting at 16:9 and switching to others on request. The options shown are 16:9 (original), 9:16 described as both speakers stacked, 1:1 for feed and 4:5 for portrait feed. The 9:16 example is instructive: start in 16:9 and ask for 9:16, and both speakers stay in the shot, stacked. That means social-ready vertical versions can be produced from the same source project rather than being re-cut from scratch in another tool. Overall, Narrative's approach is to treat the conversation and the timeline as one workspace rather than bolting a chat box onto a traditional editor. You upload your footage, describe the edit, and the agent reads the material — including the transcript — makes the cut, and tells you what it did. Every instruction becomes a version you can inspect or restore, and each turn reports the tools and credits it used. You decide how much horsepower to apply by choosing Fast, Balanced or Max, and you control the output shape through the frame settings. Narrative handles rendering and storage, so you are not managing exports, disk space or a render queue yourself. The result keeps a real, open timeline at the centre instead of a one-shot generated file. The stated benefits follow from that model. Edits arrive faster: a three minute highlight from two hours of wedding footage was on the timeline before a coffee was cold, and a producer can ask for the five best moments as verticals with captions and then open the timeline to check the work. Work that previously required a freelancer and a week can become a sentence and ten minutes. Because the output is a real editor rather than a fixed render, changes stay possible — when legal wants a frame changed, the frame is changed. Version history protects the cut from accidental damage, and Narrative's own launch video was made entirely in Narrative. Concrete scenarios named on the site include wedding films, podcasts, match highlights and product demos. A wedding filmmaker asked for a three minute highlight with the vows in the middle from two hours of footage. A podcast producer asks for the five best moments as verticals with captions. A football club's media lead asks for every goal in a ninety minute match. A product marketer produces unboxing demos. Broader examples include a rough cut, a supercut, or a vertical reel from a two hour shoot, plus podcast clips and launch videos. The audience is filmmakers, producers, media leads, product marketers and teams who cut regularly. Plans run from Free at $0/month with 3 AI prompts, the whole editor and your own footage, through Plus at $20/month with 2,000 AI credits, 10 projects, 100 GB of footage, 4K renders without watermark, every model tier including Max and version history kept for a year, and Pro at $40/month with 4,000 AI credits, 30 projects and 500 GB of footage, to Studio at $100/month for teams cutting every day with 10,000 AI credits, unlimited projects, unmetered renders with priority in the queue, 1 TB of footage and priority support. Paid plans include a 3-day trial and can be cancelled any time, and there is an iOS app in addition to the web editor. Narrative's value proposition is straightforward: describe the edit you want in plain words, and get a finished draft on a timeline you can still open and change. It compresses the distance between raw footage and a usable cut while keeping the edit genuinely editable, with versions, model tiers and frame controls under your hand.