# How to Vibe Code Your Own Colossyan (and Stop Paying for It)

> AI video and course creation platform for corporate learning

- Site: https://colossyan.com
- Category: AI Video & Corporate Learning
- Verdict: **Don't bother** (15/100 vibecodeable)
- Estimated effort: 6+ months of full-time work for a partial UI/script wrapper subset; full clone impossible

## Verdict

Build a simple document-to-script outline planner with AI text generation, but do not attempt to clone the custom neural video renderer or SCORM enterprise exporter.

Colossyan's value is locked behind millions of dollars in proprietary neural video rendering infrastructure (NEO 2), custom AI avatar training pipelines, and enterprise LMS compliance wrappers. While a solo developer can easily build an LLM wrapper that parses PDFs into text scripts and calls standard text-to-speech APIs, generating photorealistic talking-head video with real-time lip-sync at enterprise scale is fundamentally impossible for a solo builder without a dedicated GPU cluster.

### What you can't replicate

- Proprietary NEO 2 photorealistic avatar rendering engine
- Custom studio avatar generation pipelines with green-screen capture
- Enterprise SOC 2 Type II security posture and direct LMS vendor partnerships

## What it does

Transforms raw text, PDFs, and slide decks into structured video modules and training courses narrated by photorealistic AI avatars.

### Core features

- Document-to-video agent pipeline (parsing PDFs, URLs, PPTs into scene drafts)
- Interactive course builder with branching scenarios and quizzes
- Multi-language text translation with automated lip-syncing
- Avatar and voice management library (stock and custom clones)
- SCORM 1.2/2004 packaging and LMS export engine
- Workspace collaboration with version control and scene approval gates

## The business

### Pricing

- Starter: $27/mo
- Professional: $59/mo
- Enterprise: Custom

### Funding

$28M raised.
- Seed (July 2021): €1M (~$1.2M)
- Series A (Early 2023): $5M
- Series A Extension (February 2024): $22M
Investors: Lakestar, Launchub Ventures, Day One Capital, Emerge Education, APX

Founded 2020.
Team size: 51–200 employees.

## The hard parts

- Photorealistic neural avatar video rendering pipeline (NEO 2) requiring immense GPU clusters
- Phoneme-level audio-visual synchronization across 100+ languages
- Strict SCORM XML manifest generation and runtime LMS API wrapper compliance
- Agentic document ingestion and citation tracking across multi-page source files

## How to vibe code Colossyan

### Prerequisites

- Node.js (free): Runtime environment for the Next.js web application stack
- GitHub (free): Source code repository and CI/CD deployment pipeline

### Recommended AI tools

- Claude Code: Terminal agent best suited for bootstrapping full-stack scaffolding and complex multi-file logic
- Cursor: AI-native code editor for iterative frontend component styling and script editor views

### Stack

- Frontend: Next.js with Tailwind CSS and shadcn/ui components
- Backend: Next.js Server Actions and API Routes
- Database: Turso (SQLite at the edge) for lightweight relational scene and user data storage
- Auth: better-auth for self-hosted TypeScript authentication
- Payments: Stripe
- Other: OpenAI API for document-to-script agent planning, ElevenLabs API for text-to-speech audio narration, Mux API for video file hosting and playback

### Hosting

- Vercel (Hosting the Next.js frontend application and serverless backend functions): $0-20/mo
- Cloudflare (Object storage (R2) for uploaded source PDFs and generated audio assets): $0-5/mo

### Build guide

1. **Scaffold Next.js Project & Configure Database** — Initialize the Next.js application, set up Tailwind CSS, configure shadcn/ui components, and connect Turso SQLite with Drizzle ORM.

```
Create a new Next.js project using App Router, TypeScript, and Tailwind CSS. Initialize shadcn/ui and configure a Drizzle ORM schema targeting Turso SQLite. Define tables for users, projects, scenes, and video_jobs. Implement better-auth for email/password authentication with full TypeScript integration and secure session handling. Ensure the layout includes a dashboard shell with a sidebar and clean typography suitable for a professional corporate SaaS interface.
```

2. **Build Document Ingestion & Agentic Script Planner** — Implement file upload functionality for PDFs and URLs, and integrate an LLM pipeline to extract content and break it into structured video scenes with source citations.

```
Build a document ingestion pipeline in Next.js using Cloudflare R2 for storing uploaded PDFs and Word documents. Implement a server action that extracts text from documents, sends chunks to the OpenAI API with structured JSON output, and generates a multi-scene video script outline. Each scene must include a title, spoken narration script, and an explicit citation reference back to the source document paragraph to support human verification.
```

3. **Create Scene Editor & Timeline Interface** — Develop a multi-scene timeline editor where users can reorder scenes, edit scripts, select mock avatar placeholders, and apply brand kit colors.

```
Build an interactive scene editor dashboard component with drag-and-drop reordering (using framer-motion or dnd-kit). Display a vertical sequence of scene cards, each containing a thumbnail placeholder, an editable script textarea, an avatar selector dropdown, and brand color pickers. Implement real-time state synchronization with server actions so edits auto-save as users refine their video drafts before approval.
```

4. **Integrate Text-to-Speech & Video Assembly Pipeline** — Connect the ElevenLabs API to convert scene scripts into audio files and build a mock video rendering queue that compiles audio and static slide frames into a preview video using Mux.

```
Implement an audio generation service using the ElevenLabs API that converts approved scene scripts into MP3 voiceovers. Create a background worker endpoint using Inngest that stitches scene audio tracks and background images together, uploads the resulting composite video file to Mux for streaming playback, and updates the database job status from 'draft' to 'rendered'. Handle API error states and rate limits gracefully.
```

5. **Build Interactive Course Builder & Quizzes** — Add a module builder for inserting knowledge checks, multiple-choice quizzes, and branching scenario nodes between video scenes.

```
Build a course authoring module called 'Colossyan Learn' allowing users to group multiple video scenes into structured training chapters. Create a quiz builder interface enabling creators to add multiple-choice questions, knowledge checks, and branching decision points at the end of specific video modules. Store quiz structures in Turso and build a student-facing player view that pauses video playback when a check-in triggers.
```

6. **Implement Localization & Basic Export** — Add automated translation hooks for script localization across multiple languages and a basic export utility.

```
Implement a script translation feature using the OpenAI API to translate scene scripts into selected target languages (e.g., Spanish, French, German). Build an export settings modal that allows users to download the compiled MP4 video from Mux and generate a basic downloadable metadata package. Ensure all localized strings update correctly in the timeline editor state.
```

### Cost vs paying

**Starting costs (one-time):**

- Custom Domain: $12/yr
- AI API Credits (OpenAI & ElevenLabs): $30 one-time
- Total: ~$42 one-time

**Ongoing costs (monthly):**

- Vercel Hobby/Pro Hosting: $0-20/mo
- Mux & Cloudflare Storage Usage: ~$10/mo
- Total: ~$30/mo

- Paying for the SaaS instead: $59/mo
- Build time: 60–80 hours
- AI tool credits: $20/mo (Claude/Cursor Pro)
- Break-even: N/A (Personal learning clone; commercial replication impossible)

## Sources

- [Colossyan Official Website & Pricing](https://colossyan.com)
- [Colossyan Press & Series A Funding Announcement](https://www.colossyan.com/press)