Key Takeaways
- Eleven pipelines are listed in the README, including explainers, cinematic trailers, documentary montages, talking heads and localization and dubbing
- Every production stage from proposal to publish can pause for your sign-off, according to the README
- Free paths exist: stock and archive footage, offline Piper narration and, with a GPU, local video models
- Setup is a developer task: Python, FFmpeg, Node.js and a coding assistant account
- The desktop "Studio" on the website is a separate product, still in private alpha
📍 In One Sentence
OpenMontage is an open-source pipeline that lets an AI coding assistant plan, generate and render complete videos, with approval gates and budget limits along the way.
💬 In Plain Terms
You open the project in your coding assistant and describe the video you want. The assistant follows a fixed production workflow, asks you to approve the script and scene plan, calls the image, video and voice tools you have connected, and renders the result on your computer.
What Is OpenMontage?
OpenMontage calls itself "the first open-source, agentic video production system". It is a Python and Node.js project of pipeline definitions, tools and instruction files; an AI coding assistant reads those instructions and carries out the work, so the assistant is the producer and OpenMontage is its toolbox. Outputs include image-based videos, animated explainers, 3D and motion-graphics pieces, and real-footage montages cut from free stock and archive sources such as Pexels, Archive.org, NASA and Wikimedia Commons.
- Pipelines: eleven are tabulated in the README, from animated explainers and cinematic trailers to clip factories and podcast repurposing
- Research first: the agent searches the web and writes a cited research brief before it writes the script
- Tool registry: new tools and pipelines are added as Python files and YAML manifests that the registry discovers on its own
- Output formats: 16:9, 1:1 and 9:16, with word-level captions burned in
- Studio: a desktop app of the same name is announced on the website; its company is OpenMontage, Inc. and it is a different product from the repository
How OpenMontage Produces a Video
Each pipeline runs the same staged flow: research, proposal, script, scene plan, assets, edit and compose. A markdown "director skill" tells the assistant how to execute every stage, and the README says the proposal, script, scene plan, generated assets and publish steps pause for your approval.
Approval gates
- What the README says:
- A stage marked completed is rejected unless your sign-off is recorded, and superseded checkpoints are archived as an audit trail.
Budget controls
- What the README says:
- Costs are estimated before a call and reconciled after it. The default total cap is $10 and the default per-action approval threshold is $0.50; both can be changed.
Provider selection
- What the README says:
- A scoring engine ranks image, video, voice and music providers on seven dimensions, and the chosen provider and score are logged.
Quality checks
- What the README says:
- Pre-compose validation blocks plans that would produce slideshow-like video; a post-render review checks frames, audio levels and captions.
Reference video
- What the README says:
- You can paste a YouTube video, Short, Reel, TikTok or local clip, and the agent proposes two or three concepts with cost estimates.
These controls are described in the README; this review did not run a production to confirm that each one behaves as stated.
How Much Does OpenMontage Cost?
The OpenMontage software is free; the money goes to the AI assistant and to any cloud media providers you connect. Every API key in the .env file is optional, and the cheapest route uses free stock footage plus offline narration.
OpenMontage repository
- Cost:
- Free, AGPL-3.0 licence
Coding assistant
- Cost:
- Your own Claude, Cursor, Copilot or Codex plan or API usage; not included
Stock footage and images
- Cost:
- Pexels, Pixabay and Unsplash keys are free, per the README
Cloud generation
- Cost:
- Billed by each provider; the README cites roughly $1.33, about $4 and about $5 for three showcase films
Local generation
- Cost:
- No per-clip fee, but it needs your own GPU hardware and electricity
Studio desktop app
- Cost:
- Planned membership "from $10 a month" with 1,000 credits, with final terms to follow in the beta
The showcase costs are the project's own figures for its demo films, not a quote for your video. The Studio price comes from the website's facts page and is not a live offer.
Install OpenMontage
OpenMontage has no installer: you clone the repository and run a setup command, so a terminal and a developer environment are required. The Releases page is empty, and the Studio desktop app is not yet available outside its private alpha.
Prerequisites
- Details:
- Python 3.10 or later, FFmpeg, Node.js 18 or later and an AI coding assistant such as Claude Code, Cursor, Copilot, Windsurf or Codex
Install
- Details:
git clone https://github.com/calesthio/OpenMontage.git, thencd OpenMontageandmake setup
Run
- Details:
- Open the folder in your coding assistant and describe the video; optionally copy
.env.exampleto.envand add provider keys
GPU option
- Details:
make install-gpu, then setVIDEO_GEN_LOCAL_ENABLED=trueand a local model in.env
Source code
- Details:
- calesthio/OpenMontage on GitHub
Website
- Details:
- openmontage.video
Privacy policy
- Details:
- OpenMontage privacy policy
Latest version
- Details:
- No version number; last push 16 October 2026
On Windows, the README says to run npx --yes npm install if npm install fails with ERR_INVALID_ARG_TYPE. The privacy policy covers the website and Studio, not the open-source repository.
Local vs. Cloud: What Runs Where
OpenMontage is hybrid: editing and rendering run on your computer, while the coding assistant and most media generation are cloud services unless you choose local tools. The README lists Windows PowerShell, macOS and Linux setup commands and states no RAM or VRAM requirement.
Coding assistant
- Where it runs:
- Cloud by default; the README lists Ollama and LM Studio support as "coming soon"
Composition and render
- Where it runs:
- Locally, with Remotion, HyperFrames and FFmpeg on your machine
Narration
- Where it runs:
- Offline with Piper, or cloud voices such as ElevenLabs, OpenAI or Google
Images and video
- Where it runs:
- Cloud providers, or local GPU models: Wan 2.2 and 2.1, Hunyuan 1.5, LTX-2 and CogVideo
Studio app
- Where it runs:
- Runs locally per the website; credit-funded work uses OpenAI and fal, and Studio sends account and device identifiers while open
The Studio data-flow statements come from the website privacy policy, updated 14 October 2026. This review did not inspect network traffic from either the repository or Studio.
Who Should Use OpenMontage?
OpenMontage suits developers and technical creators who already work inside an AI coding assistant and want a controllable, auditable video workflow. It is a poor fit if you want to double-click an app and be done.
Use OpenMontage if
- You already pay for a coding assistant and are comfortable with Python, Node.js and a terminal
- You want approval gates and spending limits before any paid generation runs
- You want a free route built on stock footage, archive footage and offline narration
Avoid OpenMontage if
- You want a desktop installer today; the Studio app is invite-only
- You need to run everything offline, including the assistant; local-LLM support is still listed as coming soon
- You plan to embed the code in closed-source software; AGPL-3.0 is a copyleft licence, so read the LICENSE and take legal advice first
- You need independent quality evidence; the showcase films are the project's own and this review did not test it
What This Review Could Not Verify
These points could not be confirmed from the repository or the website, and they are stated here once.
- No production was run, so video quality, speed and real cost are untested
- The counts differ by source: the repository description says 12 pipelines and 100+ tools, the README says 10+ pipelines and tabulates 11
- No version number exists, so no release can be compared against this review
- The website facts page does not state a licence; the AGPL-3.0 licence comes from the repository
- Whether the Studio desktop app will be free, paid or credit-based after launch is not settled
- The website mentions no operating system for Studio, while its facts page says the private alpha runs on Windows
Competitors and Alternatives
No other reviewed tool is an agent-driven video studio, so these are the closest reviewed video-generation tools. They generate clips locally; OpenMontage plans, scripts and edits whole videos around generated or stock clips.
Tool | Platforms | Price | Best For |
|---|---|---|---|
| OpenMontage | Windows/macOS/Linux | Free, AGPL-3.0 | Agent-run finished videos |
| ComfyUI | Windows/macOS/Linux | Free, GPL-3.0 | Node-based generation workflows |
| AnimateDiff | Windows/macOS/Linux | Free, Apache 2.0 | Clips from text prompts |
| Locally Uncensored | Windows/Linux | Free, optional cloud | One-installer chat and media |
Pick OpenMontage when you want the whole production managed by an assistant; pick ComfyUI or AnimateDiff when you only need to generate clips.
Frequently Asked Questions
Is OpenMontage the same as OpenMontage Studio?
No. The repository is the open-source toolkit; Studio is a desktop app from OpenMontage, Inc. in private alpha, and its facts page describes it as a separate product.
Can I use OpenMontage without a coding assistant?
The README describes the assistant as the production driver and offers no standalone interface, so you need one of the listed assistants.
Does OpenMontage send my prompts to the cloud?
The assistant you choose receives your prompts, and each connected provider receives the requests sent to it. The Studio privacy policy says prompts, files and outputs are not used to train AI models.
Can OpenMontage make video without paid generation APIs?
The README says yes for real-footage montages from free stock and archive sources, and for local GPU generation; results were not tested here.
Which operating systems does it support?
The README gives setup commands for macOS, Linux and Windows PowerShell.