Show HN: Fast agentic video editing in the browser (Tella)
Tella is all-in-one video creation. You can record, edit and publish with one tool. Today we're shipping:
1. Tella Agent: An in-app video editing agent 2. Tella Motion: Rich motion graphics support 3. Auto Capture: Agentic screen recording
Our in-app agent uses the same tools as our MCP. All year we've been exposing more of our editor to the MCP. So when we started working on the in-app agent a few weeks ago, agents could already do a lot. Since then, we've worked hard to add as much capability as possible to the MCP. So whether you're editing on tella.com or from Claude Code you get the same capabilities.
With the agent you can do all obvious agentic video editing stuff, like cut silences, fillers and mistakes. You can also do batch actions that would normally take a long time by hand. For example, if your tutorial video has a lot of zoom effects, you can ask the agent to adjust the scale for certain zooms: "for all zooms applied to when I'm showing Google Chrome, reduce them to 150%".
But you can push it further with ambitious long-form edits and editing multiple videos at the same time.
LLMs have proven to be very good at motion graphics. Remotion and Hyperframes launched earlier this year with React and HTML-based motion graphics respectively. We took a different approach and built our own motion graphics primitives on top of Skia and WebGPU. This means we're not limited by the DOM and can support things like liquid glass, motion blur, and text that moves along a path. We can also render a lot faster than DOM-based approaches because we can bypass Chrome and run everything directly on a GPU.
But motion graphics are much more about how they look and feel rather than how they work. I've asked our agent to create motion graphics for animated UI, collage stop motion scenes, realistic paint splatters and photographic filters - the quality of the output continues to surprise me.
Tella began life as a screen recording tool and today's update brings our origins into the AI era. Auto Capture is our new agentic screen recording tool. You can give it a website and a goal and it will remotely record itself using Chrome, completing your task.
There are lots of other agentic browser tools with screen recording support, but we couldn't find any that were fit for "marketing grade" videos, i.e. recordings done in 4K at 60fps. So we built our own.
For this to be practical (i.e. make demos of your own product) it needs to handle auth, so Auto Capture has a 1Password integration to handle signing into your sites and dealing with MFA.
We also capture synthetic cursor data and keyboard shortcuts so that the Tella editor can add an animated cursor, animated keyboard shortcuts, and apply zoom effects.
You can try all of this in your browser or connect to our MCP.
Let me know what you think.
Announcement + video here: https://x.com/9ranty/status/2115797158519550360