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.
Replies
PopPop AI Vocal Remover
Wow, this tool sounds amazing. Is there any free code to try it?
Clipto
Readdy
For video editors, searching years of footage without remembering the exact filename is probably enough reason to try this.
Clipto
@wenjun_shi Exactly. You shouldn’t have to remember where a clip lives, or even when you shot it. Just describe what you remember about the moment and let Clipto find it. Thanks Wenjun!
Not an editor, but searching old meetings without remembering the date sounds genuinely useful. Nice launch.
Clipto
@sandy_liusy Exactly! Meetings are a big use case for us too. You shouldn’t need to remember when a conversation happened, just what you remember from it. Describe the topic or moment, and Clipto can find it across your meeting history. Thanks Sandy!
Pokecut
Would love to see a longer version of the demo with a massive real-world media library.
Clipto
@anthony_cai Thanks Anthony, we tested on longer videos too and works pretty smoothly, actually we provided a demo library of 1000 royalty free clips so you can try it out on your agents as well. You can download it here: https://www.clipto.com/mcp/demo-library#footage-packs
Treblle
Damm… this is good @henrykang
Clipto
@savan_kharod1 Thanks bro! Give it a try and would love to see what crazy things your agents might create :)
Clipto
Acti
This could be useful for research too, especially when your source material is mostly video and audio rather than documents.
Clipto
@bioyue Absolutely. A lot of valuable knowledge is trapped in video and audio, while most AI research workflows still work best with text. Making that media searchable and understandable to agents opens up some really interesting research workflows. Thanks Bio!
Acti
The interesting part for me is that it’s not just searching filenames. It actually understands what’s inside the videos.
Clipto
@yjbdr Exactly. That’s the key difference for us. The filename tells you almost nothing about a video. Clipto understands the people, dialogue, scenes, and moments inside it, so agents can work with the actual content, not just the files. Thanks Junbo!
Surgeflow
The “find this moment in my footage” use case is probably the first thing I’d try.
Clipto
@new_user___1282025165cc92287e7a197 That’s probably the best place to start :) Just describe whatever you remember about the moment, even if you have no idea which file it’s in. Thanks Yiyao!
Clipto
Being able to search across local video and audio through natural language could save a ridiculous amount of time.
Clipto
Invoko
The demo where the agent finds the relevant material and keeps working from it was probably the most interesting part.
Clipto
@mingyouagi Exactly. Finding the right material is only the first step. Once the agent can understand and pull from your media library, it can keep working instead of stopping at a list of search results. That’s where it gets really interesting—when the world’s most capable AI meets your own media and memories, entirely new possibilities start to open up. Thanks!
@mingyouagi Totally agree! That’s the exciting part for us too — the agent can find the right material and turn it into a video. Thanks for checking it out and for the support!