🧠 Pensieve - Context-Aware Memory for Obsidian

Pensieve is a smart memory assistant for Obsidian that helps you never lose track of your thoughts, projects, or ideas. It continuously distills your raw notes, journals, and project logs into concise, structured summaries — and helps you pick up right where you left off, even after a long break.

Whether you're researching, coding, or writing, Pensieve captures your mental context, highlights key insights, and transforms rough notes into clean, shareable ideas using the LLM of your choice — including local models.

Stay organized without slowing down. Let your notes think with you.

The Pensieve dashboard showing project overview and memory insights

Caution

Early Release: This plugin is still under active development. Features may change, and you may encounter bugs or incomplete functionality. But your notes are still safe because pensieve follows the principal of adding without removing.. so it nver touches your original only create new ones on top of them

✨ Features

🧭 Smart Memory Management

• Project Memory Tracking: Each project gets its own memory system with tasks, insights, and status tracking
• Context-Aware Summaries: Automatically extracts key information from your notes using AI
• Memory Checkpoints: Never lose track of where you left off on any project

🧹 Note Refinement Pipeline

• Clean Note Generation: Transform messy notes into well-structured, shareable content
• Configurable Output: Control tone, verbosity, and formatting
• Batch Processing: Process multiple notes efficiently with compression ratios
• Internal Linking: Automatic generation of Obsidian wiki-links

📊 Interactive Dashboard

• Project Overview: Visual status of all your projects at a glance
• Memory Insights: View recent insights, tasks, and blockers
• Autolog Analysis: Analyze patterns across your daily/weekly/monthly activities
• Real-time Updates: See memory updates as they happen

🤖 Multi-LLM Support

• OpenAI: GPT-3.5, GPT-4, GPT-4 Turbo
• Google Gemini: Gemini Pro, Gemini 1.5, Gemini 2.0
• Anthropic: Claude 3, Claude 3.5, Claude 2
• Local Models: Ollama integration for privacy-focused users
• Custom Endpoints: Support for any compatible API

⚡ Automation & Scheduling

• Automatic Updates: Set intervals for memory updates, note refinement, and autolog generation
• Smart Filtering: Exclude specific projects from automation
• Background Processing: Runs without interrupting your workflow

🚀 Quick Start

1. Installation

1. Manual Installation:

   • Download the latest release
   • Extract to your Obsidian plugins folder: {vault}/.obsidian/plugins/pensieve/
   • Enable the plugin in Obsidian settings

2. From Obsidian Community Plugins (Coming Soon):

   • Open Obsidian Settings → Community Plugins
   • Search for "Pensieve"
   • Install and enable

2. Initial Setup

1. Configure Folders:

   • Set your Journal Folder (where daily notes are stored)
   • Set your Project Root (where all project folders are located)
   • Set your Memory Folder (where memory files will be stored)
   • Set your Clean Output Folder (where refined notes will be saved)

2. Configure LLM Provider:

   • Choose your preferred LLM provider (OpenAI, Gemini, Anthropic, or Ollama)
   • Enter your API key (if required)
   • Select your preferred model

3. Enable Features:

   • Toggle automatic memory updates
   • Configure note refinement settings
   • Set up autolog generation

Complete settings panel with all configuration options

3. First Run

1. Open the Dashboard:

   • Use the command palette: Pensieve: Open Dashboard
   • Or click the brain icon in the ribbon

2. Generate Initial Memories:

   • Use the dashboard to run your first memory update
   • Review the generated insights and tasks

3. Refine Your Notes:

   • Run note refinement on a project
   • Check the output in your Clean Output folder

📁 Folder Structure

Pensieve creates the following structure in your vault:

Your Vault/
├── Journal/                    # Your daily notes
├── Projects/                   # Your project folders
│   ├── Project1/
│   ├── Project2/
│   └── ...
├── Memory/                     # Generated memory files
│   ├── projects/
│   │   ├── Project1/
│   │   │   ├── tasks.md
│   │   │   ├── status.md
│   │   │   └── insights.md
│   │   └── Project2/
│   └── autologs/
│       ├── 2024-12-15-daily-autolog.md
│       └── ...
└── CleanNotes/                 # Refined notes
    ├── Project1/
    │   ├── introduction.md
    │   ├── checkpoint.md
    │   └── project-tree.md
    └── Project2/

🎯 Use Cases

🧑‍💻 For Developers

• Track project progress across multiple repositories
• Maintain context when switching between projects
• Generate clean documentation from development notes
• Identify blockers and technical debt

📚 For Researchers

• Organize research notes and findings
• Track progress across multiple research threads
• Generate summaries for papers and presentations
• Maintain context across long research cycles

✍️ For Writers

• Organize writing projects and drafts
• Track ideas and inspiration
• Generate clean versions for sharing
• Maintain context across multiple writing projects

🎓 For Students

• Organize study notes and research
• Track progress on assignments and projects
• Generate summaries for revision
• Maintain context across courses

⚙️ Configuration

LLM Provider Settings

OpenAI

• API Key: Your OpenAI API key
• Model: GPT-3.5-turbo, GPT-4, GPT-4-turbo-preview
• Rate Limits: Respects OpenAI's rate limits

Google Gemini

• API Key: Your Google AI API key
• Model: Gemini Pro, Gemini 1.5, Gemini 2.0
• Features: Supports longer context windows

Anthropic

• API Key: Your Anthropic API key
• Model: Claude 3, Claude 3.5, Claude 2
• Features: Excellent reasoning and analysis capabilities

Ollama (Local)

• URL: Defaults to http://localhost:11434
• Model: Any model available in your Ollama installation
• Privacy: All processing happens locally

Note Refinement Settings

• Compression Ratio: How many input notes to compress into one output (0.3 = 3:1)
• Verbosity: Concise, Detailed, or Comprehensive output
• Tone: Professional, Storytelling, or Sarcastic
• Emojification: Enable/disable emojis in output

Automation Settings

• Auto-update Project Memories: Daily, weekly, or monthly
• Auto-refine Notes: Automatic note refinement
• Auto-generate Autologs: Daily, weekly, or monthly summaries
• Exclude Projects: Skip specific projects from automation

Note refinement configuration panel

🎮 Commands

Core Commands

• Pensieve: Open Dashboard - Open the main dashboard
• Pensieve: Update Project Memories - Update memories for all projects
• Pensieve: Refine Project Notes - Refine notes for all projects
• Pensieve: Generate Autolog - Generate a new autolog

Project-Specific Commands

• Pensieve: Update Memory for Specific Project - Update memory for one project
• Pensieve: Refine Notes for Specific Project - Refine notes for one project
• Pensieve: Show Project Context - Show recent context for a project

Utility Commands

• Pensieve: Export Insights as Note - Export dashboard insights to a note
• Pensieve: Test LLM Connection - Test your LLM provider connection

Available commands in the Obsidian command palette

📊 Dashboard Guide

Autologs View

• Date Selector: Choose a specific date to view autologs
• Last N Autologs: Analyze patterns across multiple autologs
• Generate New: Create a new autolog for any date
• Analyze: Use custom prompts to analyze autolog patterns

Project Memories View

• Project Selector: Choose which project to view
• Last N Checkpoints: View recent memory updates
• Search: Filter projects by name
• Load Data: Refresh project memory data

Dashboard Panels

• Status: Current project status and progress
• Tasks: Active and completed tasks
• Insights: Recent insights, blockers, and achievements
• Summary: High-level project overview

🔧 Troubleshooting

Common Issues

LLM Connection Problems

• Check API Key: Verify your API key is correct
• Check Model: Ensure the model name is valid
• Check Network: Verify internet connection for cloud providers
• Check Ollama: Ensure Ollama is running for local models

No Notes Processed

• Check Folders: Verify folder paths in settings
• Check File Types: Ensure notes are markdown files
• Check Permissions: Verify Obsidian has access to your vault

Poor Quality Output

• Adjust Settings: Try different verbosity or tone settings
• Check Input: Ensure source notes have sufficient content
• Try Different Model: Some models work better for specific tasks

Memory Not Updating

• Check Automation: Verify automatic updates are enabled
• Check Exclusions: Ensure projects aren't in the exclude list
• Check Scheduler: Verify the scheduler is running

Getting Help

1. Check the Logs: Open Developer Tools (F12) and check the console
2. Review Settings: Double-check all configuration settings
3. Test Manually: Try running commands manually to isolate issues
4. Report Issues: Create an issue on GitHub with detailed information

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Development Setup

1. Clone the repository
2. Install dependencies: npm install
3. Build the plugin: npm run build
4. Copy to your Obsidian plugins folder: npm run copy

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Obsidian Team for the amazing platform
• OpenAI, Google, Anthropic, and Ollama for their LLM services
• Community Contributors for feedback and improvements

📞 Support

• GitHub Issues: Report bugs and request features
• Discussions: Ask questions and share ideas
• Documentation: Check the Developer Guide for technical details

---

Made with ❤️ for the Obsidian community

Pensieve - Because your thoughts deserve to be remembered.