OpenMontage
OpenMontage is an open-source video-production project that a coding assistant operates through files and executable tools. Its repository describes a process covering research, scripting, asset creation, editing and composition. You supply a brief; an assistant capable of reading the project and running code follows the production stages.
Choose a production path before generating assets
The project lists pipelines for explainers, documentary montages, talking-head edits, podcast clips and other formats. A documentary path can search a corpus of real footage from sources such as open archives, while other paths use generated images or video. Choosing that path determines the material and tools the assistant needs.
OpenMontage can also analyze a reference video to propose concepts, costs and a sample. A reference is useful for discussing pacing or structure, but it does not grant permission to reproduce another creator's material. Inspect the sources and usage terms for footage, music and other assets you plan to publish.
The local Backlot board displays scripts, scenes and production decisions. The repository describes scene-by-scene approval before rendering, as well as checkpoints for resuming work. Use those approval points to reject inaccurate visuals or a poor cut before spending more on production.
Understand the local setup and optional providers
The quick start requires Python 3.10 or later, FFmpeg, Node.js 18 or later and an AI coding assistant. The project supports rendering through Remotion, HyperFrames or FFmpeg; HyperFrames has a separate Node.js 22-or-later requirement. This is a technical setup, rather than a hosted editor that works immediately after signup.
The README describes offline Piper narration and local composition tools, plus footage from open archives. Some stock providers require free developer keys. Optional paid providers add generation and other capabilities, and local GPU models are another documented path. A zero-paid-key media workflow does not mean that your chosen coding assistant or computer has no cost.
Review the output and the running budget
The repository describes cost estimates, spend caps and per-action approvals. Check the selected provider and budget before generation; demonstration costs are examples for those productions, not a fixed price for your brief.
Its post-render checks include file validation, sampled frames, audio levels and subtitles. These are useful checks for technical defects, but a successful check does not establish that the script is accurate or the video is well edited. Watch the full result before release.
OpenMontage lists the GNU AGPLv3 license. Review that license and the separate terms for models and source media before adapting the project for a commercial service or delivering a finished production.