What it is
Hypit gives AI agents such as Claude Code and Codex a language and system for creating video. You drop in a reference video and the agent clones it as a complete workflow of footage, captions, B-roll and effects, all anchored to words instead of seconds. You can also start from templates or describe a video and let the agent write the workflow from scratch. Generation models are optional, because a workflow can compile captions, motion graphics and code-rendered visuals without calling one.
Who it's for
- Developers using coding agents (Claude Code, Codex) to produce video
- Marketers and creators who want many variants of a proven video format, such as ads, UGC, podcast clips and interview clips
- Teams that need localized versions of one video in multiple languages
- Contributors who want to build and share reusable video components
Requirements
Requirements
- A coding agent such as Claude Code or Codex
- Node.js 22.15+ (per the README badge)
- Accounts for your coding agent and model services, which have their own charges
- A model service: HypiHub is recommended, or your own API or local models
Setup
Install the Hypit skill
Installs the Skill globally for your agent. On first use, the agent checks for the Hypit executable and helps prepare it if needed. Your video project can live anywhere.
bashnpx skills add hypit-ai/hypit -g
Examples
Clone a reference video with new content
Prompt/hypit Clone this video: /path/to/video.mp4, and replace the ranking content with a comparison of Hypit (official website: hypit.ai) with other AI video products.Expected output: Starts a session in a project directory and asks the agent to clone a local video while swapping the ranking content.
Create a video without a reference
Prompt/hypit Make a ranking video that puts Hypit in S tier.Expected output: Shows that cloning is optional: the agent writes the workflow from a description alone.
Pros & cons
Pros
- Pro:Produces an editable, re-runnable composition rather than a one-off render, so the same workflow can yield many variants
- Pro:Open source with no Hypit seat pricing, per-render fees or added watermark
- Pro:Works with a model service of your choice and can render code-driven visuals without generation API calls
- Pro:Documented examples (UGC, podcast, street interview) with stated total costs of about $1.07 to $1.15 each
Cons
- Con:Requires a separate coding agent and model services, each with its own accounts and charges
- Con:Licensed under a Hypit Open Source License described as Apache-2.0 with conditions, so the terms need review
- Con:Marketing claims like "100M views" are not backed by documented results in the README
Images
