AI Video Editing: 10 Open-Source GitHub Repos, Checked
A viral post lists 10 open-source repos that turn hours of editing into minutes. We checked each on GitHub: licenses, last releases, which run on a GPU, which your coding agent can drive, and where to start.
On October 16, 2026 @heyrimsha listed "10 powerful GitHub repos every creator should bookmark" for AI video editing: chat with your timeline, cut by transcript, clip podcasts into shorts, lip-sync dubbed audio.
We checked all ten on GitHub: license, activity, and what each one needs to run. The code of all ten is open source, and three joined AgentGid's video assembly catalog today. One has gone quiet, two carry licenses that matter for commercial work, and only two are built to be driven by a coding agent.
The original post by @heyrimsha on X (October 16, 2026)
Three kinds of tool on one list
The ten, checked
| Repo | License | Last release | Note |
|---|---|---|---|
| OpenChatCut | AGPL-3.0 | v0.2.17, Oct 15, 2026 | Created in July; renders with Remotion |
| FireRed-OpenStoryline | Apache-2.0 | none published | Last commit July 31, 2026 |
| AutoClip | MIT | v1.5.6, Oct 9, 2026 | In our catalog |
| Herk Edit Kit | MIT (its LICENSE file) | none published | Skills for Claude Code and Codex, built on HyperFrames and GSAP |
| AutoSubs | MIT | v3.11.0, Sep 29, 2026 | In our catalog |
| StoryToolkitAI | GPL-3.0 | v0.25.1, Feb 20, 2025 | Slowing down |
| FunClip | MIT | v2.2.1, Sep 1, 2026 | In our catalog; speech recognition by FunASR |
| NarratoAI | MIT | v0.8.7, Jul 17, 2026 | Its GitHub description is in Chinese |
| BackgroundRemover | MIT | v0.4.5, Jul 10, 2026 | Command line, images and video |
| LatentSync | Apache-2.0 code; model weights openrail++ | none published | Last commit June 20, 2025 |
Release dates and licenses from GitHub's API and each repository, read October 17, 2026.
What to watch for
- OpenChatCut is AGPL. Fine for editing your own videos. If you change it and offer it to others over a network, you must publish your changes. It also renders with Remotion, whose own license asks companies of more than 3 employees to buy a company license (see our Remotion entry).
- LatentSync's weights are not plain open source. The code is Apache-2.0, but the model on Hugging Face is under openrail++, a license with use restrictions. Its last code commit is from June 2025.
- StoryToolkitAI has not shipped a release since February 2025, and FireRed-OpenStoryline has never published one; both still get commits.
- Lip sync needs a real GPU, as the post itself says about LatentSync.
Stars are not the same as use
If you edit with Claude Code or Codex
Herk Edit Kit is the one built for that: skills that cut dead air, find mistakes, add captions, place B-roll and build motion graphics, on top of HyperFrames and GSAP. GSAP is free to use but its license bans no-code animation builders that compete with Webflow; see our motion design guide for the details and for the skills that keep an agent's animation code current.
For a pipeline you own end to end, AgentGid's Short-video kit turns a script into a vertical video with voiceover, stock footage and captions, and keeps a record of every license. Its sibling, the transcription kit, makes the transcript and subtitles that tools like FunClip and AutoSubs cut from.
Where to start
- Podcasts and interviews into shorts: AutoClip, or FunClip if you want to pick the lines yourself from the transcript.
- Subtitles inside your editor: AutoSubs for DaVinci Resolve, Premiere Pro and After Effects, running on your machine.
- Editing with your coding agent: Herk Edit Kit, or OpenChatCut if you want a timeline with agents built in.
- Lip sync for dubbed video: LatentSync, after reading the openrail++ terms of the weights.
We did not run these tools; everything above comes from the repositories, their license files and GitHub's API.