Meet Omni — an all-in-one copilot that can help you with everyday tasks like conversations, decision-making, coding, and image generation.
Key Features • Installation • How To Use • Credits • Website Link • License
- Multi-Model AI Chat
- Chat with multiple large language models and select the best model for each task.
- Get help with questions, brainstorming, writing, decision-making, and more.
- Image Generation
- Create high-quality images from natural-language descriptions.
- Resume Tailoring
- Optimize a resume for a specific role by matching it to the job description.
- Quiz Generation
- Automatically generate custom quizzes from uploaded documents.
- Create questions manually for greater control.
- Flashcard Generation
- Generate AI-powered flashcards for studying and review.
- Create and organize flashcards manually.
- File Analysis
- Upload documents for the AI to analyze, summarize, or transform.
- AI-Assisted Coding
- Get coding guidance, suggested solutions, and help debugging errors.
- Conversation History
- Save and revisit previous conversations with Omni.
- User Accounts
- Securely register, sign in, and manage personalized AI sessions.
Before getting started, ensure you have:
- Node.js installed
- npm
- A Supabase project
- OpenAI and/or DeepSeek API keys
- An Anthropic API key (resume tailoring)
- A pdflatex compile host, or a local TeX Live / MiKTeX install (see below)
-
Clone the repository:
git clone https://github.com/ryan1337c/Omni.git cd Omni -
Install dependencies:
npm install
-
Create a
.env.localfile in the project root:NEXT_PUBLIC_SUPABASE_URL=your_supabase_project_url NEXT_PUBLIC_ANON_KEY=your_supabase_anon_key SUPABASE_SERVICE_ROLE_KEY=your_supabase_service_role_key OPENAI_API_KEY=your_openai_api_key DEEPSEEK_API_KEY=your_deepseek_api_key ANTHROPIC_API_KEY=your_anthropic_api_key LATEX_COMPILE_URL=http://localhost:8080 # Optional: skip the HTTP compile service and spawn local pdflatex instead # LATEX_BIN=pdflatex CLIENT_ID=your_google_oauth_client_id CLIENT_SECRET=your_google_oauth_client_secret EMAIL_FROM=your_email_address EMAIL_APP_PASSWORD=your_email_app_password
Google OAuth and email credentials are only required for their corresponding features.
Resume tailoring calls Claude Sonnet 5, then compiles Jake’s Resume LaTeX with pdflatex. Vercel serverless cannot ship TeX Live, so production must set
LATEX_COMPILE_URLto a latex-on-http compatible host (origin is enough; the app appends/builds/sync). Locally you can run that container and pointLATEX_COMPILE_URLathttp://localhost:8080, or setLATEX_BINto apdflatexbinary instead.Resume content is PII. A public compile host sees the full document, so do not use a shared/public latex-on-http instance in production. Run a small TeX Live container you control and set
LATEX_COMPILE_URLto that host.Important: Never commit your
.env.localfile or expose secret API keys publicly. -
Start the development server:
npm run dev
-
Open http://localhost:3000 in your browser.
Create and run an optimized production build:
npm run build
npm start- Create an account or sign in.
- Choose an AI model for your task.
- Enter a prompt or upload a supported document.
- Use Omni to chat, generate images, tailor resumes, create quizzes and flashcards, or get coding assistance.
- Access previous conversations from your chat history.
Developed by Ryan Chen.
- OpenAI
- GPT-4o Mini — general chat and file analysis
- GPT-5.1 — quiz and flashcard generation
- DALL·E 3 — image generation
- DeepSeek
- DeepSeek Chat — general chat and reasoning
- Anthropic
- Claude Sonnet 5 — resume tailoring (Jake’s Resume LaTeX)
Built with Next.js, React, Tailwind CSS, and Supabase.
Copyright © 2026 Ryan Chen. All rights reserved. This project is not currently licensed for copying, modification, or distribution.