Sdk
AI video generation SDK — JSX for videos. One API for Kling, Flux, ElevenLabs, Veed. Built on Vercel AI SDK.
Last verified:
What is Sdk?
The varg SDK (vargai) is a TypeScript SDK that unifies AI video, image, voice, and music generation under one API. Instead of learning five different SDKs for different AI providers, developers learn one. It comes with a unique JSX syntax for composing videos declaratively—write <Clip>, <Image>, <Speech> and get a rendered MP4 output.
The SDK is built for AI agents and high-volume teams who need to generate thousands of consistent videos at scale. AI agents write varg-react code out of the box because the declarative API maps naturally to how LLMs think about structure. When agents encounter issues (about 5% of cases), they receive clear runtime errors with actionable messages instead of cryptic stack traces.
Key features include automatic caching (same prompt = instant cache hit at $0), parallel generation of independent elements, support for multiple AI providers (fal.ai, ElevenLabs, OpenAI Sora, Replicate, Higgsfield), and a CLI for authentication, rendering, and quick generation. The SDK uses 16 composable components like <Animate>, <Clip>, <Image>, <Video>, <Speech>, <Music>, and <Captions> to build any video pipeline from simple clips to full characters.
The tool is for developers building AI-powered video content, AI agent teams creating video autonomously, marketers producing ads, and anyone needing programmatic access to AI video generation without managing multiple provider APIs. It requires Bun or Node.js runtime with FFmpeg installed for video composition.
Sdk pricing
Pricing model: Freemium
The SDK itself costs $0 with Apache 2.0 license. You pay AI providers directly: fal.ai charges ~$0.01-0.10 per image and ~$0.50-2.00 per video. ElevenLabs has a free tier and charges ~$0.30 per 1K characters for voice. Credit packages available through varg gateway: Starter (2,000 credits for $20), up to Enterprise (100,000 credits for $1,000). Cache hits are always free - same prompt plus parameters equals $0 instant result.
Sdk pros
- One API key gives access to all AI providers
- Unique JSX syntax for declarative video composition
- Automatic caching saves money on repeated generations
- Parallel generation of independent elements
- Clear runtime errors with actionable fix hints
- Type-safe props catch mistakes early
- Supports fal.ai, ElevenLabs, OpenAI Sora, Replicate, Higgsfield
- Built specifically for AI agents to write code naturally
- CLI for authentication, rendering, and balance checking
- Apache 2.0 open-source license
- SDK itself is completely free ($0)
- Content-addressed cache survives restarts
- No hidden state or magic strings to debug
- Composable 16-component system for infinite possibilities
- Faster installs and native TypeScript with Bun runtime
Sdk cons
- Requires FFmpeg installed for video composition
- Server-side only - cannot run in browser/client components
- Does not work in Vercel Edge Runtime
- Does not work in Vercel Serverless due to timeout limits
- Video generation takes 90-180 seconds per clip
- Must pay AI providers directly for generation costs
- Requires Bun or Node.js runtime environment
- Not suitable for precise frame-by-frame motion graphics
Frequently asked questions about Sdk
What is the varg SDK?
The varg SDK (vargai) is a TypeScript SDK that unifies AI video, image, voice, and music generation under one API. Instead of learning five different SDKs, you learn one. It comes with a unique JSX syntax for composing videos declaratively—write <Clip>, <Image>, <Speech> and get a rendered MP4.
Does varg use React?
No. The JSX syntax looks like React but doesn't use React at all. It's a custom JSX runtime that transforms your components into FFmpeg render instructions. You get the familiar developer experience without the React dependency.
What runtime does varg require?
Bun is the default and recommended runtime. Node.js works too, but Bun gives you faster installs, native TypeScript, and better performance. The SDK uses Bun-specific APIs like Bun.file() but falls back gracefully.
Which AI providers are supported?
fal.ai (video, image, lipsync), ElevenLabs (voice, music), OpenAI (Sora video), Replicate (1000+ models), and Higgsfield (character generation). You only need API keys for providers you actually use.
Can I use varg in the browser?
No. The SDK requires FFmpeg for video composition, file system access for caching, and server-side AI API calls. Use it in Node.js/Bun servers, CLI tools, or serverless functions with FFmpeg layers.
How much does it cost to use varg?
The SDK itself: $0 with Apache 2.0 license. You pay AI providers directly: fal.ai ~$0.01-0.10/image, ~$0.50-2.00/video. ElevenLabs has a free tier, ~$0.30/1K characters. Caching dramatically reduces costs on iterative work.
How does varg compare to Remotion?
Different tools for different jobs. Remotion renders React frame-by-frame — great for motion graphics and data viz. Varg uses AI generation + FFmpeg composition — great for AI-generated content, talking heads, and ads. Use Remotion for precise animations, Varg for AI-powered content.
Can I use varg in Next.js?
Yes, with constraints. Works in Server Components, API Routes, and Server Actions. Does not work in Client Components, Edge Runtime, or Vercel Serverless (FFmpeg + timeout limits). Best setup: separate Node.js/Bun service for video rendering.
How long does video generation take?
Depends on what you generate. Image (Flux Schnell): 3-5s. Video (Kling 2.5): 90-180s. Voice (ElevenLabs): 2-5s. FFmpeg composition: 5-30s. Cached element: <100ms. A 30-second video might take 3-5 min first render, 10 seconds cached.