opensourceprojects.dev

A broadsheet for software that doesn't ask for your email

Paste one link, get a vertical video with subtitles and cover
GitHub RepoViews32

Project Description

View on GitHub

One Link In, One Vertical Video Out: AutoClip Takes a Run at Short-Form Editing

You've got a long interview, a podcast episode, or a conference talk that you think would do well as a short. Now you need to find the good moment, crop it to vertical, frame the speaker properly, subtitle it, design a cover, and write a caption. That's an afternoon of work per clip, if you're fast. AutoClip is an open-source desktop app that tries to collapse that into pasting a link, picking a platform, and letting the machine do the first pass.

What It Does

AutoClip is a local editing and rendering tool with a simple pitch: paste one link, choose your target platform, and it generates a video, a cover, and publishing copy. It's aimed at the short-form vertical world—Douyin, Xiaohongshu, TikTok, Reels, YouTube Shorts, Bilibili, and YouTube are all listed as supported targets. The app runs on your own computer, so the editing and rendering happen locally.

It ships in three forms: a desktop app, a CLI, and an MCP server, all updated in sync. The MCP piece is notable if you work with AI assistants—it exposes the tool to agent workflows, and there's a get_version method alongside the rest. The project is MIT licensed, and there's a version check you can run with autoclip --version to confirm what you're on.

Why It's Cool

  • It handles the whole pipeline, not just one step. Plenty of tools do transcription or auto-cropping in isolation. AutoClip's claim is the full chain—video, subtitles, cover, and copy—from a single input. That's the difference between a utility and a workflow.

  • The framing is real, not an afterthought. The README calls out "人物取景" (subject framing) as something the latest release fixed, which suggests it's actively tracking faces or speakers to keep them in frame when converting horizontal footage to vertical. That's the part most auto-clippers get wrong.

  • You can go in and fix things. The pitch isn't "trust the machine." There's an editor you can drop into when the automatic result needs adjustment. That's an honest design choice—auto-editing is never perfect, and pretending otherwise would be worse.

  • It's local-first with a honest cost model. The software is free; cloud models are billed per use. That's a clear line between what you own and what you pay for, and it means your footage isn't being uploaded wholesale to some service you don't control.

  • The MCP server is a smart bet. Making the tool callable from agent environments means it can slot into automated pipelines rather than just being a standalone GUI. For a project in this space, that's forward-looking without being hype.

  • There are real examples. The README links to actual finished clips—Bill Gates interviews, podcast segments, debate footage—with timestamps and source links. You can watch the output before you download anything, which is more than most projects offer.

How to Try It

  1. Grab the latest desktop build from the releases page. Desktop, CLI, and MCP versions ship together.
  2. Verify your install with:
    autoclip --version
    
  3. Paste a link, pick your platform, and let it generate. If the result needs work, open the editor.
  4. Check the case library to see what output looks like across styles—interview format, full-screen podcast, and so on.
  5. The quick start section of the README has the details, and the official site covers the rest.

The repo is at github.com/zhouxiaoka/autoclip. It's MIT licensed, and there are issue templates if you hit problems.

Final Thoughts

AutoClip is aimed squarely at people who publish short-form video regularly and want to cut the grunt work—creators repurposing long interviews, podcasters mining episodes for clips, anyone who needs vertical output without a timeline editor. The honest caveat is the cloud model billing: the app is free, but the intelligence behind it isn't, so heavy use has a cost. If you're comfortable with that trade and you want a local tool that gets you 80 percent of the way to a finished clip, this is worth a look. The MCP integration in particular makes it one to watch if you're building automated content pipelines.


Follow @githubprojects for more developer tools and open source projects.

Back to Projects
Project ID: autoclipLast updated: November 22, 2026 at 02:49 AM