Inspiration

Contracts are everywhere, from creator brand deals and apartment leases to gig apps, but most people skim or skip them because legalese is intimidating, especially for young users just starting out. Clausely was inspired by watching friends sign “standard” agreements that quietly hid exclusivity, auto-renewals, and vague deductions that hurt them later. The question that sparked the build: what if contracts felt like a social feed, with emojis and plain vibes, so people actually read them before signing.

What it does

Clausely lets anyone drag-and-drop a PDF and instantly translates legal clauses into Gen Z-friendly summaries with emojis , flags red flags, and drafts a polite counter-proposal email. The goal is simple: increase legal literacy the moment it matters, right before a signature, so people sign with confidence, not confusion.

How we built it

Frontend: React + Vite + Tailwind +glassmorphism UI with gradients.

Upload & parsing: drag-and-drop via react-dropzone; PDFs parsed server-side with PyPDF2 before sending text to the AI layer.

Backend: FastAPI endpoints, wrapping Gemini 2.0 with retry logic and JSON validation.

Orchestration: Pydantic with Google Generative AI bindings to handle chunking, prompting, and consistent output formatting.

Challenges we ran into

Processing pdf data, due to format error between frontend and backend. Ensuring the model always returns strict JSON (no code fences, trailing commas, or extra prose) required prompt scaffolding plus a sanitizer to clean markdown fences before json.loads.

Accomplishments that we're proud of

Shipped an end-to-end AI pipeline that ingests PDFs, extracts clause text and returns structured JSON with risks associated with each clause and Gen Z translations using Gemini 2.0 Flash.

What we learned

Pydantic Framework

What's next for Clausely

To include a chat-style Q&A: “Is this fair?” “Should I sign?”. So users can ask follow-up questions and get concise, culturally fluent answers. Supabase for storing analyses by user/session .

Built With

  • fastapi
  • googlegemini
  • pydantic
  • python
  • react
  • render
  • tailwind
  • vercel
Share this project:

Updates