Veridia

Structuring AI to help you learn.

Overview

Veridia is an AI-powered STEM learning workspace designed to support real academic workflows, not just generate answers. It helps users ask better questions, receive structured explanations, draft lab reports, and create clean graphs inside a focused interface built around clarity and usability.

Screenshots

Home Page

Application Interface

Features

• AI-powered STEM problem solving across subjects such as math, physics, chemistry, statistics, and related technical topics
• Structured, readable explanations designed to feel guided rather than generic
• Dedicated graphing workflow with downloadable outputs
• File uploads for prompt and lab workflows, including PDF, image, and text-based context
• Threaded AI conversations with recent and history views for continuity
• Lab Helper workflow for turning rough notes into stronger report drafts
• Feedback submission flow for product iteration
• Admin controls for user and feedback review
• Clean responsive interface with light and dark mode support

Tech Stack

• Frontend: Next.js 15, React 19, TypeScript, Tailwind CSS, Framer Motion
• Backend: FastAPI, SQLAlchemy, PostgreSQL, Python
• AI: OpenAI API
• Analytics: PostHog
• Graphing and parsing: matplotlib, NumPy, PyPDF, Pillow
• Billing and auth: Stripe, Google Identity / Google Auth

How It Works

1. A user submits a prompt, lab draft, or graphing request from the Veridia workspace.
2. The backend processes the request, attaches any uploaded context, and routes the task through the appropriate service.
3. OpenAI-generated output is structured and returned to the frontend for readable presentation.
4. For graphing and file-based workflows, Veridia can generate downloadable visual outputs or use uploaded materials to ground the response.

Current Status

Public Beta

Veridia is currently in a public beta phase. The product is actively being refined through testing, iteration, and user feedback, and features may continue to evolve as the platform matures.

Deployment

Frontend on Vercel

• Set NEXT_PUBLIC_API_URL to the public Render backend origin, for example https://veridia-api.onrender.com
• Set NEXT_PUBLIC_GOOGLE_CLIENT_ID
• Set NEXT_PUBLIC_STRIPE_PRICE_ID
• Set NEXT_PUBLIC_BETA_FREE_MODE

Backend on Render

• Start command: python -m app.main
• The backend reads Render's runtime PORT automatically and binds to 0.0.0.0 outside development
• Set ENVIRONMENT=production
• Set APP_URL to the public Vercel frontend origin
• Set API_URL to the public Render backend origin
• Set SECRET_KEY to a strong random value
• Set DATABASE_URL to the externally hosted Postgres connection string
• Set BACKEND_CORS_ORIGINS to a comma-separated list of allowed frontend origins
• Set GOOGLE_CLIENT_ID
• Set GOOGLE_CLIENT_SECRET
• Set OPENAI_API_KEY
• Set STRIPE_SECRET_KEY
• Set STRIPE_WEBHOOK_SECRET
• Set STRIPE_PRICE_ID

Session Cookies

• For Vercel-to-Render deployments, use SESSION_COOKIE_SECURE=true and SESSION_COOKIE_SAMESITE=none
• Leave SESSION_COOKIE_DOMAIN unset unless you intentionally need a shared parent-domain cookie
• Local development can continue using SESSION_COOKIE_SECURE=false and SESSION_COOKIE_SAMESITE=lax

File Storage

• Veridia currently treats uploaded files and generated graphs as beta-stage filesystem assets stored under STORAGE_ROOT
• The default backend storage path is local and therefore ephemeral on Render unless you attach a persistent disk
• To use a Render disk later, point STORAGE_ROOT at the mounted disk path; UPLOAD_DIR and GRAPHS_DIR can stay under that root or be overridden explicitly