Clipto
Fully local, natural language search over terabytes of media
Start new thread

Clipto MCP - Let agents source clips from terabytes of your local video

by
•
Clipto MCP gives Claude, ChatGPT, and other AI agents the ability to source clips and more from inside the videos, photos, and audio recordings stored on your computer. Instead of manually browsing files, simply describe what you need. For example, turn a script into a video by matching each sentence with your local footage; find every scene where someone mentioned a topic; create rough cuts; or search years of media as if you have a dedicated assistant editor.

Add a comment

Replies

Best

I can see this fitting really naturally into an editing workflow. Find the right clips first, then build from there.

Maker

 Exactly. Finding the right clips is often the hardest part. Once agents can understand and source from your entire media library, there’s a lot more they can help with downstream. Thanks Lily!

Does this work the same way with Claude and Cursor?

Maker

 yes, it is the same. Just download the Clipto app and install the MCP for the corresponding agent.

I’d be interested to see this with a really messy archive — random filenames, old projects, mixed audio and video.

Maker

 That’s the kind of library Clipto is built for. Filenames and folder structures only help when you already know what you’re looking for. Clipto looks at what’s actually inside the files, so even a messy mix of old projects, audio, and video can become searchable through natural language.

Maker
Yes, because humans don’t think in file names. We remember people, places, opinions, and moments.

As someone with TONS of footage, not the best file organization (to be kind to myself about it) and content creation as a side project, I am really excited to try this one, congrats on the launch!

Maker

 It sounds like Clipto could be a great fit for your workflow. You shouldn’t need a perfect folder structure—or spend your limited side-project time organizing files—before you can use your footage. Just describe the moment you remember and let Clipto find it.

You can also use our invite code to get a one-month trial. We’d love for you to give it a try and share any feedback or suggestions. Thanks so much!

Maker
Feel free to give it a try — sign up with code PHLNCH to get one month free. Would love to hear what you think!

Claude and Cursor can already read local files. Why do I need Clipto MCP in between? What does it actually add?

Maker
To be completely honest, if you give an agent a few files directly, it can do something similar. The difference becomes clear when you’re working with terabytes of local media. At that scale, three things really matter: time, privacy, and cost. Clipto analyzes and indexes your media locally and relatively quickly, without consuming any additional tokens. This is especially valuable for media professionals with years of video and audio. Sending that much content to an online AI could take longer than you’re willing to wait—and the bill could be enormous

 That makes sense. But when I search my local media through Claude or Cursor, what actually leaves my Mac? Are the original files uploaded, or does the Agent only receive the search results?

Maker

 Absolutely right. Frankly, if your workflow consists of a few specific and relatively short videos, these agents can handle it perfectly well—but when you're dealing with a lot of large files locally, Clipto comes in handy.

Maker
Clipto only sends the abstracted results produced by the local analysis, not the original media itself. If you’re only using search, the AI never sees or accesses the source video or audio.
It’s pretty mind blowing that it can work with local files like this. Is the multimodal model understanding the content? If that’s the case, I could feed my local videos directly into it for editing. That would really boost my efficiency. Honestly, this would’ve been unthinkable just a year ago!
Maker

 Yes, that’s the key. Clipto uses multimodal models to understand what’s actually inside your local video and audio, not just filenames or transcripts.

It’s amazing how far AI has come in just one year. We’re working hard to turn that progress into something genuinely useful for real workflows like this. Once your media is indexed locally, MCP lets you bring that understanding into the AI tools you already use, opening up a lot more possibilities. Thanks!

Maker
Definitely not easy! Took us almost two years to get here 😄

Sourcing clips from local video terabytes using AI agents is a huge time-saver. Congrats on bringing this to life!

Maker

 Thanks! That’s exactly the idea. Once a media library grows into terabytes, finding the right clip can take longer than editing it. We want agents to understand your entire library and bring back the right moment in seconds.

I like the new use case for the application! Because photos are easier to store than ever, I find it harder to find what I'm looking for in my camera roll than before.

Maker

 Exactly. The easier it becomes to capture and store everything, the harder it becomes to find a specific moment later. We want you to simply describe what you remember and let Clipto find it for you. Thanks sean!

Maker
It’s not just about generating anymore — it’s about finding!

2tb indexed in a day is wild 🤯 local search finaly usable. agencies your main users?

Maker

 Thanks! Agencies are definitely a strong fit, especially when they have years of footage spread across different clients and projects. But we also see independent creators, editors, podcasters, and researchers using it.

The common thread isn’t company size—it’s having enough media that folders and filenames stop being useful.

Maker
This tool is also in high demand among in-house marketing and social teams with large media archives. We were recently approached by a company with 2TB of footage that needed a faster way to find the right moments for short-form social content.

The fact that it can go from searching your media to actually using what it finds is the part I want to explore more.

Maker
We think the real leap is from “find it for me” to “use it for me.” The Elon video we made is a fun example — MCP didn’t just help find the relevant media, but actually used what it found to help create something new. That’s where it gets really interesting.
First
Previous
123
•••
Next