Technical Overview & Integration Notes

Architecture Explainpaper operates as a hosted SaaS web application. Uploaded PDFs are parsed server-side and fed into an LLM context window, enabling the model to answer questions and generate explanations grounded in the full document.

Core Feature Modules

• Highlight Explanations: Selected text spans are sent to the LLM along with surrounding paper context, returning a simplified explanation. Complexity level is a user-controlled parameter.
• Paper Chat: A retrieval-augmented or full-context chat interface allowing multi-turn Q&A with the paper. Responses include citations to specific sections.
• Auto-Insights: Server-side generation of structured outlines, key point extraction, and concept relationship mapping upon paper upload.

Model Tiers

• Free tier: Basic AI model (likely GPT-3.5 class or equivalent)
• Pro tier: Advanced AI model (likely GPT-4 class or equivalent), required for full-paper summarization

Language Support Explanations can be returned in 50+ languages, suggesting a multilingual LLM backbone with language detection or user-preference routing.

No public API is currently documented. Integration is via the web UI only. Highlights and explanations are persisted in user accounts on the Pro plan.