AI Organiser: Intelligent Note Organization for Obsidian

A comprehensive AI-powered plugin for Obsidian with 37+ commands across tagging, summarization, research, meeting minutes, presentations, semantic search, and more. Supports 14 cloud providers and 5+ local LLM options.

Installation

From Community Plugins (Recommended)

1. Open Obsidian Settings > Community Plugins > Browse
2. Search for AI Organiser
3. Click Install, then Enable

Manual Installation

1. Download main.js, manifest.json, and styles.css from the latest release
2. Create folder <vault>/.obsidian/plugins/ai-organiser/
3. Copy the three files into the folder
4. Reload Obsidian and enable the plugin in Settings > Community Plugins

Quick Start

1. Open Settings > AI Organiser > AI Provider
2. Select a provider (Claude, OpenAI, Gemini, Groq, DeepSeek, etc.)
3. Enter your API key
4. Open the Command Picker (configurable hotkey) to browse all commands

Features

AI Chat

Full-featured chat with 6 modes:

Mode	Purpose
Note	Ask questions about the active note
Vault	RAG-powered search across your knowledge base
Highlight	Discuss selected text or highlights
Research	Web research with source quality scoring
Free Chat	Open conversation with file attachments and projects
Slides	Build themed HTML presentations with PPTX export

Features: conversation persistence, project memory, global memory, smart document indexing (ONNX RAG), model switching, and resume from previous conversations.

Presentation Builder (Slides Mode)

• Describe your presentation in natural language
• LLM generates rich HTML slides with themed CSS
• Preview in sandboxed iframe with slide navigation
• Chat to refine individual slides or the whole deck
• Export to editable PPTX (via dom-to-pptx) or self-contained HTML
• Optional brand guidelines integration with Haiku audit
• Version history with undo/redo

Intelligent Tagging

• Taxonomy-based tagging with customizable themes and disciplines
• 3-tier hierarchical tags (e.g., science/biology/genetics)
• Multiple modes: generate new, match existing, hybrid, or custom
• Batch operations for folders or entire vault
• Tag network visualization with D3.js interactive graph
• Taxonomy guardrail validates tags against your taxonomy

Content Summarization

• URLs - Web articles with link preservation
• PDFs - Native multimodal support (Claude/Gemini) or text extraction
• YouTube - Via captions extraction
• Audio - Transcription + summarization (MP3, WAV, M4A, OGG, WebM)
• Multi-source - Summarize multiple sources into one note
• 5 built-in personas - Student, Executive, Casual, Researcher, Technical
• Custom personas via config file
• Reviewed Edits - Inline diff preview before any write

Web Research Assistant

• Multi-provider search (Claude Web Search, Tavily, Bright Data)
• Smart escalation: free fetch > Web Unlocker > Scraping Browser
• Source quality scoring (5 weighted signals)
• Academic mode with DOI extraction and citation formatting
• Multi-perspective query decomposition
• Usage guardrails with monthly budget tracking
• Zotero integration for citation management
• Vault pre-check (search vault before web)
• Streaming synthesis with citations

Meeting Minutes

• Generate structured minutes from audio transcripts
• GTD overlay: classify actions by context (@office, @call, etc.)
• 2 built-in personas (Standard, Governance) + custom personas
• Terminology dictionaries for transcription accuracy
• Context document support (agendas, presentations)
• Chunked processing for long meetings (5000-token chunks)
• Word document (DOCX) export
• Obsidian Tasks format support

Audio Recording & Transcription

• In-plugin audio recording (desktop + mobile)
• Whisper-compatible transcription
• Post-transcription cleanup (keep/compress/delete)
• Direct integration with minutes and summarization

Smart Digitisation

• Extract text from handwritten notes, whiteboards, diagrams
• 5 modes: auto, handwriting, diagram, whiteboard, mixed
• Built-in sketch pad with pressure-sensitive drawing
• Image compression with backlink-safe vault replacement

Kindle Sync

• Import from My Clippings.txt or Amazon cloud
• Differential sync (only new highlights)
• 4 highlight styles: blockquote, callout, bullet, plain
• Color grouping and cover images
• AI auto-tagging after import

Newsletter Digest

• Fetch unread Gmail newsletters via Google Apps Script
• AI triage summaries per newsletter
• Rolling daily digest notes
• Key links extraction (spam-filtered)
• Auto-fetch scheduler

Semantic Search & RAG

• 7 embedding providers (OpenAI, Gemini, Ollama, Cohere, Voyage AI, OpenRouter, local ONNX)
• Vector store with chunk-based indexing
• Related notes sidebar (auto-updates on note switch)
• RAG-enhanced AI responses
• Multi-select export from search results
• Zero-setup ONNX fallback (no API key needed)

Canvas Toolkit

• Investigation Board - RAG-based related note visualization
• Context Board - Embedded content mapping (YouTube, PDF, web, audio)
• Cluster Board - Tag-based note grouping with LLM clustering

Mermaid Chat

• Conversational diagram editing with live preview
• Version history and line-level diff view
• Template library (built-in + custom)
• Multi-format export (SVG, PNG, .mermaid, canvas)
• Staleness detection (warns when note content changes after diagram creation)
• Type conversion between 12 diagram types

Document Export

• Export to PDF, Word (DOCX), or PowerPoint (PPTX)
• 5 color scheme presets + custom colors
• Configurable font family and size
• Flashcard export (Anki, Brainscape)
• Meeting minutes DOCX with structured sections

Vault Tools

• Quick Peek - Fast 1-paragraph triage of embedded sources
• Web Reader - Batch URL triage and note creation
• Find Embeds - Vault hygiene scan for assets and orphans
• Pending Integration - Auto-resolve embedded content before integration
• Obsidian Bases - Structured metadata and dashboard generation

Translation

• Translate notes, selections, or external sources
• 20+ target languages
• Smart dispatch: selection > multi-source > note
• Multi-source translation (URLs, YouTube, PDFs, documents, audio)

Supported Providers

Cloud LLM Providers

Provider	Models
Anthropic (Claude)	Claude 4.6, Sonnet 4.6, Haiku 4.5
OpenAI	GPT-5.4, GPT-4o, o3, o4-mini
Google (Gemini)	Gemini 2.5 Pro/Flash
DeepSeek	DeepSeek R1, V3
Groq	Llama, Mixtral (fast inference)
Mistral	Mistral Large, Codestral
Cohere	Command R+
Grok	Grok-2
OpenRouter	200+ models
Requesty	Multi-provider routing
Amazon Bedrock	Claude, Llama, Titan
Google Vertex AI	Gemini models
Alibaba (Aliyun)	Qwen models
SiliconFlow	Open-source models

Local LLM Options

• Ollama - Run models locally
• LM Studio - GUI-based local inference
• LocalAI - OpenAI-compatible local server
• Jan - Desktop LLM app
• KoboldCpp - GGUF model runner

Embedding Providers

OpenAI, Gemini, Ollama, Cohere, Voyage AI, OpenRouter, Local ONNX (zero-setup)

Mobile Support

Works on iOS and Android with:

• Cloud-only provider mode
• RAM-aware indexing limits
• Touch-optimized modals
• Automatic audio codec negotiation

Internationalization

Full interface and output language support for English and Simplified Chinese. Output language configurable independently from interface language.

Configuration

All settings are organized into collapsible sections:

• AI Provider - Main LLM setup
• Specialist Providers - Dedicated providers for YouTube, PDF, Audio, Flashcards
• Tagging - Tag generation and taxonomy
• Summarization - Personas, styles, transcript options
• Capture & Input - Audio, digitisation, sketch pad
• Meeting Minutes - Output, timezone, personas, GTD
• Vault Intelligence - Semantic search, canvas, RAG
• Integrations - Bases, NotebookLM, document export
• Preferences - Language, interface, mobile
• Advanced - Configuration files

Contributing

1. Clone the repo: git clone https://github.com/Lbstrydom/ai-organiser.git
2. Install dependencies: npm install
3. Development build: npm run dev
4. Production build: npm run build
5. Run tests: npm test

See AGENTS.md for detailed architecture documentation.

License

MIT License - see LICENSE for details.

Support

If you find this plugin useful: