What Is Rendervid? The Agent-Native Video Render Engine, Explained (2026)
If you searched "what is Rendervid," first a quick disambiguation, because the name is crowded: this is Rendervid, the AI-agent video render engine from QualityUnit, not Render.com the cloud host, and not the unrelated C++ video library of a similar name. With that out of the way: Rendervid turns JSON templates into video, and its whole pitch is that AI agents can drive it directly. Here is the straight explanation.
TL;DR — What is Rendervid?
| Question | Answer |
|---|---|
| What is it? | A stateless engine that renders video from JSON templates and React |
| Who makes it? | QualityUnit (the company behind FlowHunt) |
| License / cost | Free for personal and commercial use, attribution required |
| Agent interface | Built-in MCP integration, so Claude Code, Cursor, and others can render |
| Output | MP4, WebM, MOV, plus image and GIF formats |
| Closest comparison | Remotion (which renders from React code) |
| Not for | People who want a finished clip without authoring JSON |
Rendervid is part of the video-as-code family: deterministic video produced from code rather than a timeline.
What Rendervid actually is
Rendervid is a stateless, cross-platform rendering engine. You give it a JSON template, the structure of the video, its layers, text, timing, and animations, and it renders that to a video file. It also accepts custom React components for cases the templates do not cover, and it binds your content into templates with simple {{variable}} placeholders. "Stateless" means each render is a self-contained job, which is exactly what you want when an automated agent is firing renders without a session to manage.
It is made by QualityUnit, the company behind FlowHunt and other SaaS products, so this is a company-backed project rather than a lone open-source experiment. That matters mostly for expectations: it is actively developed and documented, but it is also young.
The agent-native part
The reason Rendervid shows up in "AI video coding" conversations is its built-in MCP integration. MCP is the protocol that lets AI assistants call external tools, and Rendervid ships a server with a dozen-plus tools grouped into rendering (render a video or image, start an async render, check status), validation and discovery (validate a template, list capabilities and examples), and documentation (look up components, animations, and easing). Connect it to Claude Code, Cursor, Windsurf, or another MCP-compatible agent, and the agent can author a template, validate it, and render the video without you leaving the chat.