Local RAG Chatbot for Contract Documents

## Project Overview I'm looking for an experienced AI/LLM developer to build a local-first contract review and question-answering application using Retrieval-Augmented Generation (RAG). The system should allow users to upload contract documents, ask natural language questions about them, and receive accurate answers with document citations. This project is intended as an MVP/Proof of Concept and must run entirely on local machines, with no cloud hosting required. ## Requirements ### 1. Document Upload & Processing * Support uploading: * PDF files * DOCX files * TXT files * Parse and extract document content * Chunk documents appropriately for retrieval * Generate embeddings and store them in a local vector database * Support adding new agreements/documents after initial setup ### 2. Chat Interface Create a clean browser-based interface where users can ask questions such as: * "What is the termination clause in the ABC Vendor Agreement?" * "Which contracts include a non-compete clause?" * "What are the payment terms in this agreement?" The system should: * Return accurate answers based on document content * Cite the source document and relevant section/page whenever possible * Allow users to: * Search across all uploaded documents * Restrict queries to a specific document ### 3. Local-First Architecture Requirements: * Runs entirely on a local machine * No cloud hosting required for this phase * Data must remain local * The only external communication allowed is OpenAI API calls for answer generation (and optionally embeddings) ### Preferred Technology Stack #### LLM * OpenAI GPT-4o or GPT-4o-mini #### Embeddings * OpenAI text-embedding-3-small * OR Sentence Transformers (developer recommendation welcome) #### Vector Database * FAISS (preferred) #### Backend * Python * FastAPI (preferred) or Flask #### Frontend * Simple HTML/CSS/JavaScript * OR lightweight React application #### Document Parsing * PyPDF2 * pdfplumber * python-docx I'm open to recommendations as long as the solution remains local-first and uses the OpenAI API. --- ## Deliverables * Complete source code * Clean and organized repository (GitHub or ZIP) * README with setup instructions * Documentation for: * Installation * Configuration * Running the application locally * Adding new documents --- ## What I Will Provide * OpenAI API key(s) * Sample agreement documents for testing (NDAs, service agreements, vendor contracts, etc.) * Prompt feedback throughout development * Clear requirements and timely communication ## Ideal Freelancer * Experience building RAG applications * Strong knowledge of OpenAI APIs * Experience with vector databases (especially FAISS) * Experience with document ingestion and retrieval systems * Python/FastAPI development experience --- ## Timeline * MVP / Prototype Delivery: 1–2 weeks --- ## To Apply, Please Include 1. Links to similar RAG, document search, or document Q&A projects you've built (GitHub, demos, portfolio, etc.) 2. Your preferred technology stack for this project 3. Estimated number of days required to deliver the MVP 4. Any questions or suggestions regarding the requirements