Introducing Vidroot

Short-form video products usually stop at generation. They can create a clip, add captions, and maybe export a file. What they do not give teams is the operating system around the clip.

Vidroot is being built around that missing layer:

• organizations, members, and roles
• a durable project model for scripts, scenes, assets, and renders
• upload intents for source footage and brand kits
• orchestration hooks for downstream rendering and publishing

That gives us a better starting point for a real product team. We can support a web app now, keep a separate API for mobile later, and still share the same data model across both.

Why the architecture matters

AI video workflows are not one request. They are pipelines. Script generation, voiceovers, assets, captions, rendering, and publishing all have different latency and failure modes.

That is why the first version of Vidroot is being built around:

• Better Auth with organizations
• shared Drizzle schema for auth and app data
• R2-backed asset storage
• upload intent tracking
• a dedicated API worker for future mobile clients

The result is slower to start than a demo, but much more defensible once real usage shows up.