Nomi — Direct AI models to create videos on a timeline, using your own models and keys.
Analyzed by Sai Pavan Gopularam · AI · Video Generation · View on GitHub
- Stars: 544
- Forks: 123
- Commits last 30 days: 100
- Health: Active (100 commits this month)
- Language: TypeScript
- License: AGPL-3.0
What It Is
Imagine having a personal film director who understands your vision for a video. Nomi is a desktop application that acts like this director, letting you tell an AI agent what kind of video you want. It then plans shots, generates images and video clips using AI models you connect, and arranges them on a timeline for you to edit.
This matters because commercial AI video tools often charge high markups for models you could access cheaply yourself. Nomi lets you use models at their real cost, keeping your projects, prompts, and keys private on your machine. It solves the problem of expensive, opaque, and cloud-dependent AI video creation.
License Verdict
AGPL-3.0 License — Use with Caution — Copyleft Restrictions Apply • SaaS Use Requires Source Disclosure
The AGPL-3.0 license requires that if you modify and distribute this software, or if users interact with your modified version over a network (like in a SaaS offering), you must make the full source code of your entire application available to those users. This can restrict building proprietary commercial products directly on top of Nomi. However, you can offer professional services like custom builds or integrations, as mentioned by the project.
How to Use It
Nomi is a desktop application. For end-users, download the appropriate installer (macOS or Windows) from the releases page, then connect an AI model by pasting your key. Describe your desired video to the AI agent, then direct and export.
Prerequisites:
- macOS or Windows x64
- Node.js 20+
- pnpm
Estimated setup time: 10 minutes.
git clone https://github.com/aqm857886159/Nomi.git
cd Nomi
corepack enable
pnpm install
pnpm devWhat I'd Build With This
Nomi Skill Pack & Workflow Template Marketplace (micro-saas)
Build and sell specialized "skill packs" or advanced workflow templates for Nomi users. These could be pre-configured storyboards, character reference cards, or custom agent instructions that leverage specific models for niche video styles (e.g., explainer videos, product demos, social media shorts). Users pay for these digital assets to enhance their Nomi experience.
Effort: 1 Week Build Time · Target: Indie filmmakers, marketing agencies, content creators using Nomi · Pricing: $19-$99 per pack/template
Managed Nomi AI Video Production Service (saas)
Offer a full-service AI video production agency where Nomi is your primary internal tool. Clients provide briefs, and your team uses Nomi's powerful agent-driven workflow to generate high-quality, cost-effective videos. You manage the models, keys, and editing, delivering finished MP4s. The client interacts with your service, not Nomi directly, avoiding AGPL distribution issues for the client-facing product.
Effort: 1 Month Build Time · Target: Small businesses, marketing departments, YouTubers needing bulk content · Pricing: $500-$5000 per video project or retainer for ongoing content
Nomi Enterprise Integration & Customization (enterprise)
Provide consulting and development services to large media companies or marketing firms. This includes custom building Nomi to integrate with existing internal systems (e.g., asset management, proprietary AI models, brand guidelines), deploying it in an AGPL-compliant manner within their infrastructure, and offering ongoing support and feature development tailored to their specific needs. This leverages the "For Teams" offering.
Effort: 3 Months Build Time · Target: Large media enterprises, advertising agencies, corporate training departments · Pricing: $10,000 - $100,000+ per project or annual support contract
Sai Pavan Gopularam's Take
Nomi is a fantastic example of a local-first AI app that cuts out middlemen. The AGPL license means you can't easily build a proprietary SaaS on it, but offering enterprise integration services around it, as the project suggests, could easily net $50k+ per client project.
Watch Out For
- AGPL-3.0 License: Building proprietary software that links to or uses Nomi as a network service requires you to release your entire application's source code under AGPL-3.0. This is a significant restriction for many commercial projects.
- Unsigned macOS/Windows Builds: The desktop application installers for macOS and Windows are not signed, meaning users will encounter security warnings (Gatekeeper, SmartScreen) requiring manual overrides during the first launch.
- Local-First Only: Nomi stores all projects and data locally on your machine. There's no built-in cloud sync or collaborative features, which might be a limitation for teams or users needing to work across multiple devices.
I break down trending repos like Nomi every week — join the newsletter.