When I'm building a tool for video professionals, the most valuable feedback comes from people who actually do the work. Not hypothetical use cases. Real projects, real deadlines, real pain points.
Recently I had the chance to demo Vindexr for Dustin Samplawski, a Director of Photography based in Portland whose work spans commercial productions, documentaries, and corporate video. His portfolio includes everything from narrative brand films to run-and-gun event coverage, and he brings the perspective of someone who has to manage footage across varied projects and client expectations.
I recorded our conversation because I thought it might be useful for others considering whether Vindexr fits their workflow. The full demo is embedded below, but here are the highlights.
The setup
Before our call, Dustin had already imported footage from a two-day conference shoot. About ten presenters, each speaking for roughly 45 minutes. Hours of material that he'd already delivered as straightforward edits, but wanted to explore further for potential documentary or commercial applications.
This is exactly the scenario Vindexr is built for: large volumes of footage where you know valuable content exists, but finding specific moments manually would take hours.
Walking through the core features
We started with the basics. When you import footage into Vindexr, it automatically transcribes any audio and indexes the visual content. The transcript stays synced to the video player, so clicking any line of text jumps the playhead to that exact moment.
Dustin had already seen this working on his conference footage. The next question was what to do with it.
I showed him how to manually pull sound bites using keyboard shortcuts (I and O for in and out points, C to add to a collection) and how collections can span multiple source files. If you're building a piece that draws from several interviews, you can gather clips from across your library into a single collection.
The AI integration
The feature that got the strongest reaction was the Claude AI integration. Instead of manually searching or scrubbing, you can describe what you're looking for in natural language.
I demonstrated with footage from an entrepreneurship retreat where the speaker gave tips about succeeding at conferences. Rather than scanning through an hour of material, I typed: "Find the parts where Stu talks about tips for finding success at a conference."
Within moments, Vindexr returned 18 relevant clips with text summaries of each. One click created a collection ready for review or export.
Dustin's response: "That's amazing. I wish I had this for the project I just did."
A real-world use case: finding introductions
We talked through a specific problem both myself and Dustin have encountered. When editing interview-heavy projects, we often need to add name and title graphics for each subject. If the subjects don't introduce themselves in a consistent way, finding those moments means scrubbing through every interview to locate where each person states their name and role.
I showed him how to expand the search scope to an entire folder of interviews and prompt: "Find all the places where people introduce themselves."
The result was a collection of just those moments, each with a summary showing the person's name and title. What would have been a tedious manual process became a few seconds of typing and waiting.
"It kind of gets you 75% of the way there," Dustin said. "And then it's just adding the frosting to the cake."
DaVinci Resolve integration
Dustin uses both Premiere and Resolve, with Resolve being his primary editor lately. I walked him through the direct integration: with Resolve running, you can send any collection directly from Vindexr. It creates a bin with the referenced media and places the clips on a timeline in order.
For workflows that involve pulling selects and building rough assemblies, this eliminates the manual export and import steps. You go from "here are the clips I want" to "here they are on my timeline" in one click.
The local processing advantage
We spent some time discussing why Vindexr runs locally rather than through cloud services. Dustin brought up Frame.io as a comparison point. It's great for reviews and approvals, but uploading terabytes of raw footage for search and organization is expensive, time-consuming, and often impractical.
"I have terabytes of stuff," I explained. "I can't upload all of that. It's a headache, and you always hit the storage limit."
Vindexr keeps everything on your own hardware. The AI features use the Claude API for complex queries (which requires sending transcript text, not video files), but the core indexing, transcription, and visual search all happen locally. Your footage never leaves your machine.
Looking forward
Dustin mentioned wanting to use Vindexr for upcoming documentary and commercial projects where he needs to "sift through it all and find the good stuff without having to spend hours looking."
That's the goal. Not to replace the creative work of editing, but to eliminate the tedious searching that happens before the creative work can begin.
As I told Dustin: "It's making me like video editing again, because the worst part of editing is the tedious stuff of searching. If you can get a rough cut done and then do the fun part, polishing and adding your creative flair, that's what I like doing. Not searching for what people are saying."
Watch the full demo
The complete conversation is embedded above. If you're curious about specific features or want to see the workflow in action, it's worth watching through.
And if you're dealing with similar challenges in your own work, I'd love to have you try Vindexr. Start your subscription today!
Dustin Samplawski is a Director of Photography based in Portland. You can see his work at dustinsamplawski.com