AI-native video production
The editor is an agent.
The creator stays the director.
OpenMontage turns narration and story concepts into structured video productions — coordinating scene planning, footage and image selection, audio, effects, motion, multilingual dubbing, rendering and quality control in one agent-driven workflow.
01Narrationhuman direction
→
02Scene planagent reasoning
→
03Assetssearch + provenance
→
04Editmotion + sound
→
05QA + Rendergates + verification
What it is
A production system, not a one-click slideshow.
OpenMontage is being built inside a real creator workflow. Coding agents help operate and improve the production stack while deterministic gates verify outputs before a project is considered complete.
Scene intelligence
Narration becomes a scene plan with pacing, visual intent, transitions and platform-aware composition.
Footage & assets
Real footage and source-backed assets are prioritized, tracked and checked instead of blindly filling a timeline.
Sound & motion
Music, SFX, keyframes, motion graphics, subtitles and 2D storytelling are coordinated around the narration timeline.
Multilingual output
A Supertonic-based dubbing workflow extends production into multiple languages without rebuilding the edit from scratch.
Multiple render engines
FFmpeg, Remotion and HyperFrames provide different rendering paths for documentary edits, motion design and agent-native scenes.
Verification gates
Artifacts, project state, provenance, render properties and quality checks are validated before work is marked complete.
How it works
Human direction.
Agent execution.
Machine verification.
The creator remains responsible for the story and final creative judgement. Agents handle repeatable production work, while code-based checks reduce hallucination and “looks done” failures.
Direct
Start with narration, creative rules and the intended platform.
Plan
Agents translate the story into scenes, visual intent, timing and required assets.
Assemble
Footage, images, graphics, music, SFX, dubbing and motion are coordinated into an edit.
Verify
Production gates check artifacts, sources, render properties and project state using deterministic evidence.
Render & review
The final output is rendered, inspected and kept under human control.
Built in production
Tested on a live multi-platform creator workflow.
OpenMontage is continuously changed, tested and improved while producing factual, human-centered storytelling for Facebook, Instagram, TikTok and YouTube.
Facebook 32K
Instagram 13.7K
TikTok 5K
YouTube 1K
Founder
Creator-built, not lab-built.
OpenMontage is built by Mert Gündoğdu from the workflow behind GizemLabTR. The project started from a practical question: how much of professional video production can be delegated to agents without giving up creative control?
“I direct the story. The system helps make the production decisions, executes the edit, and proves what it actually completed.”
Mert Gündoğdu · Founder / CreatorOpenMontage
Building an agent-native production layer for video.
Currently bootstrapped and developed through continuous real-world production.
mert@openmontageai.com Replace this address with your domain email before publishing.