1.1.0
What's new
- Claude support. Claude is now available both over the HTTP API and via the local
claudeCLI, bringing the provider count to six. - PDFs and images in the knowledge base. PDF text is extracted with PDF.js; images are sent to the model as images, so you can ask about diagrams and screenshots.
- Quiz and flashcard languages. Generate content in English, Korean, or both.
- Rewritten README with setup instructions, a provider/cost table, privacy notes, and screenshots.
Requirements
Requires Obsidian 1.7.2 or newer (raised from 1.4.0).
Installing
Download all assets below into <your vault>/.obsidian/plugins/multi-ai-assistant/, then reload Obsidian and enable the plugin under Settings → Community plugins.
The sprite-*.png files and pdf.worker.min.js are loaded from the plugin folder at runtime, so they must sit alongside main.js — the character avatars and PDF reading will not work without them.