Clueso MCP - Create and edit videos by chatting
by•
Make videos with Claude, ChatGPT, or any AI agent. Give Clueso an idea, deck, reference video, or recording, and it handles the whole video production: storyboard, scenes, voiceover, music, and editing, on-brand. Everything stays editable, by hand or by chat.
Replies
@OpenAI add them as an official plugin this is SO cool!!
Is it possible to use cloned voices? via Fish Audio or ElevenLabs.
Sounds good! Can I simply connect it with my claude by MCP?
@davitausberlin Absolutely! You can find the connection prompt here.
Congratulations on the launch! 🎉 Really curious about the iteration side of Clueso. Coming from a design/3D background, I`ve found that AI tools can get you surprisingly close on the first pass, but the last 10-20% detailing, often becomes a back-and-forth of approximations.
How precise would you say editing by chat gets here? If I want a specific change to one scene or frame, can Clueso reliably hit that intent, or is there still some trial and error involved?
@theo_gzd this is definitely a challenge with most AI creative tools and we have focused a lot on this at Clueso! With Clueso, everything you create stays editable via our powerful video editing interface. You can edit it manually or by prompting. There's motion graphics parameters, keyframes, very robust element properties, etc which can all be edited as needed.
Would love to hear your feedback when you get a chance to try it out!
@neel_balar I find this very useful and honestly, it feels like exactly what many artists are looking for in this evolving AI space. Having more control and faster iteration has always been essential for improving workflows, especially for smaller teams. Congrats for creating a well-designed environment where models can perform while giving users more flexibility
I will definitely come back with some feedback once I`ve had the change to test Clueso and integrate it in my workflow.
@theo_gzd awesome, thank you!
@theo_gzd likely AI? I wish I could get an AI to do what you do!
The MCP angle is what really caught my attention here. Going from “I have an idea” to a complete, editable video without having to live inside a video editor feels like a pretty meaningful shift in the workflow. Also love that you’re keeping the underlying project editable rather than treating AI output as a one-shot generation. Curious to see how people end up combining this with tools like Figma, Linear, and Slack. Congrats on the launch!
@ender_ozmen yeah, I'm extremely Claude-pilled myself, so anything I can do to not leave the terminal makes me very happy. That's how we started this off tbh but the reception to the MCP has been great. Would love to have you try it out!
@ender_ozmen Yes, we have been super bullish on letting folks edit the underlying project via our editor. That's what really helps in going from 90% there to 100% of the output you need.
Here are some workflows we have seen our users build by connecting the MCP with other tools they already use via Claude/ChatGPT:
1. Creating a new feature release video automatically whenever a feature ticket/PRD is moved to done in Jira or Linear.
2. Turning Figma prototypes into lo-fi animated product videos.
3. Tracking support queries from Slack channels and building videos automatically for most commonly asked questions.
and many other such workflows.
Would love to know what you're planning on creating!
@ender_ozmen Thanks a ton! And yes, every project created by the MCP is fully editable. This lets you review and fine tune the output to get the best version out!
Curious about the multi-frame feedback: does the agent actually get rendered frames back as images after every edit, and how do you keep that from eating the whole context on a 90-second video?
@louislecat It actually gets a SINGLE image, that contains a thumbnail of sorts for all frames it's requested for. Because most modern models are good, and a good video tends to have low fidelity (ie, there's not a lot going on screen), this actually works great in helping the model understand multiple frames in one image. This is actually how professional video editors make a storyboard as well: they combine multiple frames into one visualisable frame that you can skim through.
It doesn't get back after every edit, but only when it chooses to take a look post making edits to one clip of the video. Typically, the model will request the MCP for a dense number of frames for a specific animation or a sparse set to get a sense of the whole video.
I have been editing on Premiere and After Effects for years, so I am used to the long way of doing things.
Those tools are powerful,but editing even a simple product video meant opening the editor, cutting, writing script, voiceover, zooms, captions…easily a few hours.
Clueso is the opposite. You tell it what you want and the video is already in a decent place.
Editing with prompts is a game changer. I am not starting on a blank timeline anymore. I prompt what I need, and then I only fine-tune.
That is why it feels easier than anything else I have used.
@prathamesh_amble do you prefer editing with chat post the initial creation or do you like going into the editor post creation?
@neel_balar My first instinct is to keep editing in chat. I go into the editor only when I need a very specific fix like timing, a zoom, something small that is faster by hand. What I like is that both options are there. I don’t get stuck with only one way.
The editable project is the interesting part for me. I’d love to try making a Goldfish demo from a screen recording, then use Claude to iterate on the pacing and callouts without starting over. Congrats on the launch!
@kar_re that's a great use case for Clueso! Here's what I'd recommend: connect the MCP and then input your screen recording along with a reference video you like in terms of what you'd like the output to look like and see Clueso build the demo/launch video for you. Let me know how it goes!
This is genuinely useful.
Video creation is always one of those things that takes way longer than it should.
making this MCP-native instead of another dashboard is the smart move, since it means the video request can live in the same conversation as whatever produced the deck or recording in the first place. the part I'm curious about is scenes/storyboard revision through chat specifically: if I ask the agent to redo just one scene after the voiceover and music are already generated, does it re-render the whole video or only the piece that changed? that's usually where these AI-editing tools quietly turn a 30 second tweak into a 5 minute wait.
@neel_balar good to hear that's already the behavior rather than a roadmap item. scoped re-render is the thing that actually decides whether people iterate on tone/pacing or just accept the first take because redoing the whole thing feels expensive. will give it a real try.