SaaS Video Generator

This project is a Next.js web application that uses AI to generate educational videos. It leverages various APIs for script generation, audio synthesis, image/video generation, and rendering.

Configuration

To successfully run this project, you need to configure environment variables and authentication files.

1. Environment Variables (.env)

Create a .env file in the root directory and add the following keys:

# LLM & AI Services
NEXT_PUBLIC_LLM_API_KEY=     # Key for your primary LLM provider
GEMINI_API_KEY=              # Google Gemini API Key

# Audio Services
NEXT_PUBLIC_AUDIO_API_KEY=   # Generic Audio API Key (if used)
NEXT_PUBLIC_TRANSCRIPT_API_KEY= # Transcription Service API Key
ELEVENLABS_API_KEY=          # ElevenLabs API Key for high-quality TTS

# Google Services
GOOGLE_API_KEY=              # General Google API Key
GOOGLE_SEARCH_API_KEY=       # Custom Search JSON API Key
GOOGLE_SEARCH_CXID=          # Custom Search Engine ID (CX)

# Media Generation
FREEPIK_API_KEY=             # Freepik API Key for image/video generation

2. dynamication.json

This file (dynamication.json) controls the "personality" and logic of the content generation.

• scriptFewShot: Contains "few-shot" prompting examples (low and high complexity) used by the LLM to generate scripts. You can edit this to change the writing style of the generated videos.
• avatars: A list of available AI avatars/voices. Each entry has:
  • value: The Voice ID (e.g., ElevenLabs ID).
  • label: Display name.
  • langType: Supported language/accent.
  • gender: MALE/FEMALE.
  • languageCode: BCP-47 language tag (e.g., hi-IN).
• avatarAudioMap: Maps voice IDs to a sample audio URL for previewing voices on the frontend.

3. Google Cloud Authentication (available in mediaApis)

The directory app/mediaApis contains a Google Service Account JSON key file (e.g., axiomatic-treat-xxxx.json).

• This file is AXIOMATIC (essential/foundational) for Google Cloud services (like Vertex AI or Cloud Storage) to function.
• Ensure this file is present and correctly referenced by the code (e.g., in vertex.ts or google-cloud initialization).
• Do not commit real credentials to public repositories.

Usage

Installation

npm install

Running the Development Server

npm run dev

The application will be available at http://localhost:3000.

Video Editor / Preview

This project uses Revideo. You can start the editor to refine templates:

npm run revideo:editor

Debugging

Debug Render (debug_render_data.json)

To debug the video rendering process without regenerating all assets (which costs money and time):

1. Save Data: The system saves generation data to data/debug_render_data.json during a successful run.
2. Static Generation: You can trigger a "Static Gen" mode.
   • When the staticGen flag is set to true (see app/videoGenerator.ts), the system skips LLM/Media generation.
   • It directly reads parameters from data/debug_render_data.json.
   • It passes these parameters to renderPersonalizedVideo.

This allows you to tweak the rendering logic (React components, animations) and instantly see the result using previously generated assets.