Fastest AI Editing Workflow for Podcasts: Riverside Tutorial
A full record-to-export walkthrough of Riverside's AI editing tools, demoed live on a real two-host podcast episode about dogs.
July 3rdA hands-on test of OpenArt Director, the chat-driven "vibe directing" tool that turns one product photo into a scene-consistent, five-minute-capable AI ad.
OpenArt Director fixes AI video's two biggest flaws — short scene limits and inconsistent characters — by having you direct through chat and approve still-image previews before it renders.
OpenArt Director is a chat-driven AI video tool built to fix what breaks most AI generators: five-to-fifteen-second scene caps and characters, locations, and products that drift between clips. Instead of typing a prompt and hoping, you pick 'Guide Me' and the agent asks clarifying questions, writes a script and storyboard, and shows still-image previews of every character, location, and product before spending render credits. Trevor Jones tests it live, building a TikTok Shop-style protein powder ad from one product photo: he reviews the AI-generated model and gym, revises the look with a single chat message, approves the update, and ends with a finished nine-second vertical ad — no camera, crew, or editing skills required.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →
Trevor claims anyone with a laptop can now produce what used to require a full crew, and previews three AI-generated examples he built with OpenArt.

A single unedited prompt ("beautifully lit 15-second fight scene in the style of John Wick") produces a finished gun-fu clip with no follow-up prompting.

With a few more prompts and clarifying questions from the AI, OpenArt builds a golden-hour beach romance scene with a matching swelling score.

A one-shot TikTok Shop-style ad for a face cream comes out passable but reveals a common failure mode: one prompt changes one thing at a time and risks getting other things wrong.

Trevor names the core limitation: most tools only render 5-15 second clips and struggle to keep characters, locations, and products consistent once you stitch scenes together.

OpenArt Director lets creators build videos up to five minutes long while keeping character, location, and product consistent across every stitched scene, in any style from UGC to cinematic.

Trevor frames the pitch as "vibe directing" — the filmmaking equivalent of vibe coding — where you describe the idea and the agent asks clarifying questions instead of you guessing at prompts.

He uploads a single product photo (ProDough protein powder) into a new Director project and opens the "Vibe Direct Your Next Video" chat interface.

He types one loose prompt describing the ad and picks "Guide Me" over "Just Make It" so the agent asks clarifying questions instead of gambling on a single generation.

The agent asks for aspect ratio (9:16 for TikTok Shop) and a render-quality tier before spending any credits — he picks 720p Lite.

OpenArt generates a full concept, script, and storyboard for the spokesperson-style ad before any video is rendered, all editable at this stage.

The tool shows still images of the character, gym location, and product treatment so Trevor can approve the look before it spends the credits a full render would cost.

Unhappy with the dreary default gym and character look, he types a revision request ("make Maya blond, green eyes, higher-end gym") instead of restarting the whole project.

OpenArt regenerates just the two changed elements and shows the updated character and location for approval while keeping the product locked.

Trevor's aside: on TikTok Shop, copying the structure of ads that are already converting outperforms inventing something original, and Director makes it fast to test.

He points to the link, then plays the finished vertical ad — voiced by the AI-generated model Maya — as proof the whole pipeline holds together.
The gap between AI video tools that frustrate you and ones that work is whether you can see and fix the plan before the expensive generation step runs.
“Hollywood should be nervous.”
“some broke dude with a laptop can now create what it would have taken an entire crew, a location, a camera package, and a ton of money to create all by himself”
“It's like vibe directing, the filmmaking equivalent of vibe coding.”
“The thing that really works is to copy what's already working... it sounds like cheating, but it's actually what works.”
“She looks considerably better promoting ProDough than I ever could.”
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
Trevor opens with a dare: some "broke dude with a laptop" can now produce, solo, what used to take an entire crew, a location, and a camera package — then spends the next seven minutes trying to prove it with one tool.
The two generation modes OpenArt Director offers; Trevor recommends Guide Me for anyone who wants control before credits are spent.
Trevor's aside that mimicking proven ad structures on TikTok Shop outperforms original creative, and that OpenArt Director makes that fast to test.
“If you wanna check it out, I dropped a link in the description below and in the pinned comment.”
Soft, repeated affiliate plug (also mentioned around 2:50) rather than a single hard pitch — folded into the walkthrough instead of interrupting it.
00:00
00:07
00:13
00:17
00:24
00:29
00:35
00:42
00:47
00:50
00:55
01:02
01:04
01:11
01:16
01:21
01:25
01:32
01:37
01:44
01:50
01:51
01:58
02:06
02:08
02:13
02:17
02:23
02:29
02:35
02:40
02:46
02:51
02:56
03:01
03:07
03:12
03:16
03:23
03:26
03:33
03:38
03:44
03:49
03:54
04:00
04:05
04:10
04:15
04:21
04:26
04:31
04:34
04:43
04:47
04:52
04:57
05:03
05:08
05:13
05:19
05:24
05:29
05:32
05:41
05:45
05:50
05:53
06:02
06:06
06:11
06:17
06:22
06:27
06:35
06:38
06:43
06:48
06:54
06:59A full record-to-export walkthrough of Riverside's AI editing tools, demoed live on a real two-host podcast episode about dogs.
July 3rdA 19-minute deep-dive into the updated AI cinema pipeline — three changes that killed the plastic look for good.
May 27thA creator builds two full AI-generated websites live, connecting Claude Design to Higgsfield through one workflow rule that keeps every generated image and video visually consistent with the page around it.
August 5thA Claude skill trained on Seedance 2.5's own 35-page prompting guide turns a one-line idea into a shot-by-shot video prompt, no reading required.
August 11thA free, open-source Remotion project turns Claude Code into a director that layers AI art, AI voice, and code-driven animation into a chess lesson, a kids' story, and a paper documentary from one codebase.
August 1stA screen-recorded walkthrough of turning one long prompt — or one saved skill file — into a fully rendered Vox-style paper-collage explainer, script through voiceover, using Claude Code and the Higgsfield MCP.
July 27th