Lexio – AI-Native PDF Reader (Ollama, Claude, OpenAI, Gemini)

Name: Lexio – AI-Native PDF Reader (Ollama, Claude, OpenAI, Gemini)
Availability: InStock
Author: nikodemseb

by nikodemseb·Mar 4, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemCozy

PDF-native AI chat beats copy-paste workflow, but specialized readers crowding fast.

Strengths

•Full document context in system prompt is genuinely better UX than snippet-only chat.
•Annotation-to-comment feature tethers insights directly to PDF, not lost chat history.
•Clean provider abstraction (Ollama, Claude, OpenAI, Gemini) + 50-line extensibility.

Weaknesses

•Prototype stage (only 7 commits) means core features like save/load likely unpolished.
•PDF + AI chat combo now has multiple competitors (Adobe Acrobat AI, specialized tools).

Post Description

I built Lexio because the standard workflow is broken: copy text from your PDF, switch to a chat window, paste context, explain what you're reading, get an answer, switch back. Repeat forever.

The core idea: AI should live inside the reader, not beside it. Select any passage, hit "Ask AI", and get a response grounded in the entire document. But the feature I'm most proud of: you can summarize your entire AI conversation and attach it directly as a comment on the PDF — so your insights stay with the document, not lost in some chat history.

AI backends: - Ollama — local, fully offline, nothing leaves your machine - Claude, OpenAI, Gemini — cloud via API key - Clean provider abstraction, ~50 lines to add a new backend

Full document context is passed in the system prompt (not just the selected snippet), streaming responses throughout, multiple conversations per document.

Stack: Electron, React, PDF.js, Zustand, TypeScript.

Repo: https://github.com/nikodemseb/lexio

Happy to talk through the architecture or the annotation-to-PDF pipeline.