✦ MC PDF Studio

A polished PDF viewer CLI with dark/light themes, minimap navigation, and a multi-provider AI Sidekick (Claude, Ollama, OpenAI, xAI).

Screenshots

Install

npm install -g mc-pdf-studio

Usage

mc-pdf-studio open [file] [options]

Or without global install:

node index.js open [file] [options]

Options

| Flag | Description | Default | |------|-------------|---------| | [file] | Path to a PDF file to open immediately | — | | -p, --port <port> | Port to listen on | 3000 | | --enable-ai | Enable AI Sidekick | disabled | | --ai-provider <provider> | AI provider: anthropic, ollama, openai, xai | anthropic | | --ai-model <model> | Model name for the chosen provider | provider default | | --ai-base-url <url> | Base URL for Ollama or OpenAI-compatible providers | provider default | | --no-minimap | Hide the page minimap panel | shown | | --theme <dark\|light> | Initial color theme | dark |

Examples

Open the viewer (no file):

mc-pdf-studio open

Open a specific PDF:

mc-pdf-studio open ./report.pdf

AI Sidekick with Claude (Anthropic):

ANTHROPIC_API_KEY=sk-ant-... mc-pdf-studio open ./report.pdf --enable-ai

AI Sidekick with Ollama (local, no API key needed):

mc-pdf-studio open ./report.pdf --enable-ai --ai-provider ollama --ai-model llama3.2:latest

AI Sidekick with OpenAI:

OPENAI_API_KEY=sk-... mc-pdf-studio open ./report.pdf --enable-ai --ai-provider openai --ai-model gpt-4o

AI Sidekick with xAI (Grok):

XAI_API_KEY=... mc-pdf-studio open ./report.pdf --enable-ai --ai-provider xai --ai-model grok-beta

Light theme, custom port, no minimap:

mc-pdf-studio open ./doc.pdf --theme light --port 8080 --no-minimap

AI Provider Environment Variables

| Provider | Env Var | Notes | |----------|---------|-------| | anthropic | ANTHROPIC_API_KEY | Default provider | | ollama | — | Local, no key needed | | openai | OPENAI_API_KEY | | | xai | XAI_API_KEY | |

You can also change the provider at runtime without restarting — open Settings ⚙️ in the UI to switch providers, fetch available Ollama models, and set the active model.

Features

📄 PDF Viewer

• Renders all pages using PDF.js in the browser
• Smooth scroll through all pages
• Drag & drop PDF files to open them
• Upload button for file browsing

🗺️ Minimap Panel

• Thumbnail preview of every page in a left sidebar
• Click any thumbnail to jump to that page
• Highlights the currently visible page
• Toggle with the minimap button in the header

🔍 Navigation & Zoom

• Page number input (click to type a page number)
• Previous/Next arrow buttons
• Keyboard arrows: ←/→ or ↑/↓ to move pages
• Zoom in/out buttons and fit-to-width mode
• Keyboard: + / - / 0 for zoom

🌙 Dark / Light Theme

• Toggle with the sun/moon button in the header

✦ AI Sidekick

• Slide-in panel on the right side
• Chat about your PDF: ask questions, get summaries, extract data
• Quick-prompt chips: Summarize, Key points, Overview, Numbers & dates, Conclusions
• Full markdown rendering in responses
• Maintains conversation history within a session
• Settings UI (⚙️ button) to switch providers and models live — no restart required
• Supports: Claude by Anthropic, Ollama (local), OpenAI, xAI (Grok)

API Routes

| Route | Description | |-------|-------------| | GET /api/config | Returns server config (AI provider, theme, etc.) | | GET /api/initial-pdf | Streams the startup PDF if one was provided | | POST /api/upload | Accepts a PDF upload, returns a temp path | | GET /api/pdf?path=... | Streams an uploaded PDF by temp path | | GET /api/pdf-info?path=... | Returns page count and metadata | | POST /api/ai/chat | Chat endpoint (requires --enable-ai) | | POST /api/ai/settings | Update AI provider/model at runtime |

License

• MIT (c) Mohan Chinnappan