Overview:
Corpus is an open-source, AI-powered document Q&A platform that allows users to upload PDFs and web pages, ask questions in natural language, and receive answers with citations linking back to the source text. It is designed for developers, researchers, or anyone needing to extract answers from a collection of documents. The system organizes files into workspaces and document sets for structured management, making it suitable for knowledge retrieval tasks.
Core Features:
Natural Language Q&A with Citations: Ask questions about uploaded documents and get answers with direct citations to the source text.
Document Upload: Supports uploading PDFs and web pages for analysis.
Workspace and Document Set Organization: Group related documents into workspaces and sets for better management.
Multi-Provider LLM Support: Integrates with OpenAI (GPT-4o, GPT-4), Anthropic (Claude 4, Claude 3.5), Google (Gemini), and xAI (Grok) for answer generation.
Multi-Provider Embedding Support: Uses OpenAI and Voyage AI for document embedding and search.
Use Cases:
Researchers who need to quickly find and verify answers from a collection of uploaded PDFs.
Developers building custom knowledge base tools who require a self-hostable Q&A system with citation support.
Professionals managing project documentation who want to ask questions across organized workspaces of web pages and files.
Why It Matters:
Corpus provides a transparent, self-hostable alternative to proprietary document Q&A services. Its architecture includes a modular stack (FastAPI, PostgreSQL, Elasticsearch, RabbitMQ, Temporal, Redis, S3) and supports multiple LLM providers, allowing teams to swap AI models without platform lock-in. The citation system directly links answers to source text, offering a verifiable audit trail not always present in closed solutions.
