Cinevo AI

Cinevo AI is an AI video maker SaaS where users create, plan, edit, generate, and manage videos with an AI assistant.

Comment

Inspiration

Creating a video traditionally requires multiple steps: developing an idea, writing a script, planning scenes, creating visuals, and managing the generation process. For many creators, the hardest part is not having an idea—it is turning that idea into a structured, production-ready video.

We wanted to build a platform where users could start with nothing more than a simple idea and work with AI to transform it into a complete video plan.

That idea became Cinevo AI: an AI-powered video creation workspace designed to make video production more accessible, structured, and collaborative.

What it does

Cinevo AI allows users to go from an initial idea to a queued video-generation workflow.

Users can:

  • Create an account and securely authenticate.
  • Describe their video idea using natural language.
  • Chat with an AI assistant to develop and refine the concept.
  • Generate a structured video plan containing scenes, narration, visual prompts, audience, platform, duration, and style.
  • Edit the generated plan before starting production.
  • Queue video-generation jobs asynchronously.
  • Monitor generation progress in real time.
  • Browse, search, and filter their personal video library.
  • View individual videos, scenes, progress information, and generated output.

The platform combines AI assistance with a structured workflow instead of treating video generation as a single prompt-and-wait experience.

How we built it

Cinevo AI was designed as a full-stack application with a clear separation between the frontend, backend, AI services, database, and asynchronous processing layer.

The frontend was built with Next.js and the App Router, providing the creative workspace, authentication pages, dashboard, AI chat interface, video planning interface, and video library.

The backend was built with NestJS, exposing REST APIs and a Socket.IO gateway. Authentication uses JWT access and refresh tokens, while Prisma provides type-safe database access to PostgreSQL.

For AI capabilities, Cinevo communicates with Google Gemini through the backend. This keeps API credentials and other sensitive infrastructure secrets away from the browser.

For long-running video-generation tasks, we implemented an asynchronous job architecture using Redis and BullMQ. Instead of keeping a user's request waiting for the entire generation process, the backend creates a job that can be processed independently.

Real-time progress is delivered through Socket.IO, with HTTP polling available as a fallback. Generated outputs can be stored locally during development or through S3-compatible storage in production.

The resulting architecture is:

User
  ↓
Next.js Frontend
  ↓ REST + Socket.IO
NestJS Backend
  ├── Gemini AI
  ├── PostgreSQL + Prisma
  ├── Redis + BullMQ
  └── Video/Storage Provider

Challenges we ran into

One of the biggest challenges was designing the application around asynchronous video generation.

Video generation is not an instant operation, so a traditional request-response architecture would create poor user experiences and potentially tie up backend resources. We solved this by introducing Redis and BullMQ as a job queue, allowing generation tasks to run independently from the user's request.

Another challenge was providing a reliable real-time experience. We implemented Socket.IO events for states such as processing, progress, completion, and failure, while also supporting HTTP polling as a fallback.

Security was another important consideration. We intentionally kept Gemini credentials, database credentials, JWT secrets, Redis credentials, and storage credentials on the backend rather than exposing them to the frontend.

Finally, designing the video plan itself required thinking beyond a simple AI response. The AI output needed to become structured, editable data that the rest of the application could understand and process.

Accomplishments that we're proud of

We are proud that Cinevo AI is more than an AI chatbot or a simple video-generation interface.

We built an end-to-end product workflow:

Idea → AI Collaboration → Structured Video Plan → Editing → Generation Queue → Real-Time Progress → Video Library

Some of the accomplishments we are particularly proud of include:

  • Building a complete full-stack AI SaaS architecture.
  • Integrating Gemini into a structured creative workflow.
  • Implementing secure JWT authentication with access and refresh tokens.
  • Creating an editable scene-based video planning system.
  • Building asynchronous video-generation infrastructure with Redis and BullMQ.
  • Implementing real-time progress updates with Socket.IO.
  • Creating a searchable and filterable personal video library.
  • Keeping sensitive AI and infrastructure credentials server-side.
  • Designing the system to support scalable storage and future video providers.
  • Creating a responsive workspace that works across desktop and mobile.

What we learned

Building Cinevo AI taught us that an AI-powered application is not just about connecting an LLM to a UI.

The most important part is designing the system around the AI.

We learned how to transform unstructured natural-language ideas into structured application data, how to design asynchronous workflows for long-running tasks, and how real-time communication can significantly improve the user experience.

We also gained deeper experience with distributed application architecture by combining PostgreSQL, Redis, BullMQ, WebSockets, REST APIs, and AI services.

Most importantly, we learned to think about AI as a collaborative component of a product rather than simply a chatbot.

What's next for Cinevo AI

Cinevo AI is designed to grow into a complete AI video production platform.

Our next steps include:

  • Integrating production-grade text-to-video and image-to-video providers.
  • Supporting AI-generated voiceovers and background music.
  • Adding automatic subtitles and captions.
  • Introducing richer scene and timeline editing.
  • Adding multiple video styles and templates.
  • Supporting different aspect ratios for YouTube, TikTok, Instagram, and other platforms.
  • Adding team collaboration and shared projects.
  • Introducing usage-based credits and subscription plans.
  • Improving video-generation reliability with distributed workers.
  • Adding more advanced AI agents for scripting, visual direction, narration, and editing.

Our long-term goal is to make Cinevo AI a creative production partner where a user can bring an idea, collaborate with AI, and produce a polished video without needing a traditional production workflow.

Built With

Share this project:

Updates

Submission history