Pro Features Overview - 29.90$

Deep Ask Plugin Quick Start Guide

Purchase link: Deep Ask

📖 Introduction

Deep Ask is a powerful Obsidian AI chat plugin designed to enhance your knowledge management experience. By integrating multiple AI models, Deep Ask can understand your note content, providing intelligent question-answering, content generation, and knowledge retrieval.



🌐 Chrome Extension

Deep Ask now offers a Chrome browser extension, allowing you to seamlessly use AI assistant features while browsing the web!

Download: Chrome Web Store - DeepAsk AI

Key Features

  • Integration with Obsidian: Supports linking tags, context, and files from Obsidian to initiate Q&A on the current webpage
  • Webpage Content Conversation: Directly chat with the content of the current webpage
  • Cross-platform Collaboration: Seamlessly switch between browser and Obsidian for a complete knowledge management workflow



✨ Main Features

  • 🤖 Multi-model Support:Support GPT, Gemini, DeepSeek, Qwen, etc.
  • 💻 Real-time Content Conversation:Automatically understand your note structure and content, support conversation with current notes and webpages
  • 🏷️ Tag Integration:Context collection based on tags, intelligent content retrieval
  • 📎 File Attachment:Support adding note files as context
  • 🤖 Professional Prompt Selection:Built-in professional prompt library to enhance conversation quality
  • 🗓 Time Range Limitation:Support filtering context content by time range
  • 📷 Creative Poster Generation:AI generates HTML-formatted posters with preview and download support
  • 💡 Quick Suggestions:Intelligent recommendation of common conversation templates
  • 💬 Streaming conversation:Real-time response, support conversation history management
  • 🔄 Multi-session management:Support creating multiple conversation topics
  • 🌍 Multi-language support:Support Chinese and English interfaces
  • 🔌 MCP Server:Built-in MCP (Model Context Protocol) server, allowing AI tools like Claude Desktop and Cursor to directly access your Obsidian vault

🚀 Quick Start

Step 1: Install Prerequisites

Deep Ask depends on the Dataview plugin, please ensure it is installed and enabled:

  1. Open Obsidian settings
  2. Enter the "Community plugins" section
  3. Search and install Dataview
  4. Enable Dataview plugin

Step 2: Install Deep Ask Plugin

  1. Download the plugin file to the Obsidian plugin directory, usually in your repository's .obsidian/plugins directory
  2. Restart Obsidian
  3. Enable Deep Ask in the plugin list

Step 3: Activate the Plugin

  1. Open the plugin settings page
  2. Enter the "Activation" tab
  3. Enter your account email and activation code

⚙️ Configuration Guide

AI Model Configuration

In the "AI Model" tab of the settings page, you can configure the supported AI services:

Here is an example of configuring the DeepSeek model:

🔵 DeepSeek Configuration

Fill in the API Key, and then click the AI Model, and select the specific model to enable:

You can also customize the AI model configuration:

🟢 Custom Configuration

You can customize the AI model configuration, supporting the following parameters:

  • Base URL:The base URL of the AI service
  • API Key:The API Key of the AI service
  • AI Model:The supported models of the AI service

AI Context Configuration

  • Maximum conversation rounds:Set the number of conversation history to retain
  • Temperature:Set the randomness of AI replies, the lower the more stable, the higher the more random
  • Exclude files with matching tags:Generally used to filter periodic notes and template notes. The former has already queried flash notes in periodic notes through tags (duplicate context), and the latter are various template notes (invalid context)
  • Use current opened file only:Only use the currently opened file as context, ignore all other files opened in the background
  • LifeOS Integration Config:Configure file search strategy when using LifeOS with AI features
    • Theme Note File Search:Select how AI searches theme-related files
      • Include index files only:Only include index files
      • Include entire folder:Include entire folder
    • Recursive Sub-theme Search:Enable recursive search for sub-themes within the selected theme tags

Conversation Settings

  • Maximum conversation rounds:Set the number of conversation history to retain
  • Temperature:Set the randomness of AI replies, the lower the more stable, the higher the more random

🎯 Usage Guide

Launching the Chat Interface

There are several ways to open the AI chat:

  1. Toolbar button:Click the Deep Ask AI icon in the sidebar
  2. Command PaletteCtrl/Cmd + P search "Deep Ask AI Chat"
  3. Shortcut:Can be customized in settings

Basic Conversation

  1. Enter your question in the chat interface at the bottom
  2. Select the AI model (if multiple are configured)
  3. Click Send or press Enter to send
  4. AI will reply to your question in real-time

Advanced Features

💻 Real-time Content Conversation

  • Directly chat with the currently opened file and webpage, supporting multiple files and webpages as context, and linking with existing notes

📎 File Attachment

  • Trigger file search by typing /, select the file to add to the context
  • The opened file will be automatically added to the context

🏷️ Tag Integration

  • Trigger tag search by typing #, select the tag to add to the context
  • DeepAsk will automatically search for all records in the vault, including tasks, flash notes, and files with the selected tag

🤖 Professional Prompt Selection

  • Enter / to trigger prompt search, select the prompt to add to the context
  • DeepAsk will display a list of prompt files in the specified directory for selection

🗓 Time Range Limitation

  • Enter /2025-08-22 to trigger time range limitation
  • DeepAsk will automatically limit all matching context (files, tags) to the specified time range

📷 Creative Poster Generation

  • Require AI to generate HTML-formatted posters
  • Click the "eye" preview button below the AI reply to view and download the poster

💬 Session Management

  • Click "+" to create a new conversation
  • Each conversation manages the context independently

✏️ Message Editing

  • Hover to display the edit button
  • Support editing the sent message, triggering AI to regenerate the reply
  • Support clicking the regenerate button next to the message, triggering AI to regenerate the reply

💾 Download AI Conversation

  • Click the plus button next to the message, and the AI reply will be appended to the current note
  • Click the download button in the lower right corner of the dialog, and the entire conversation will be downloaded as a Markdown file

💡 Quick Suggestions

In the blank conversation, Deep Ask will display commonly used suggestions:

  • /2025-08-22 Summarize today's diary?
  • #SwapHouse What are the good and bad things I did?
  • #HistoryCycle Summarize this tag?

🔌 MCP Server

Deep Ask includes a built-in MCP (Model Context Protocol) server, allowing you to access your Obsidian vault directly through AI tools like Claude Desktop, Cursor, and Claude Code.

What is MCP?

MCP (Model Context Protocol) is an open protocol that allows AI applications to securely connect to external data sources and tools. With MCP, you can let AI assistants like Claude directly read and operate on your Obsidian notes.

Configuration

Claude Desktop Configuration

  1. Make sure the Deep Ask plugin is enabled and running
  2. Open the Claude Desktop configuration file:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  3. Add the following configuration:
{
  "mcpServers": {
    "obsidian-lifeos": {
      "url": "http://127.0.0.1:51377/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_ACTIVATION_CODE"
      }
    }
  }
}
  1. Restart Claude Desktop

Cursor Configuration

Add MCP server in Cursor settings:

{
  "mcpServers": {
    "obsidian-lifeos": {
      "url": "http://127.0.0.1:51377/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_ACTIVATION_CODE"
      }
    }
  }
}

Claude Code Configuration

Create a .mcp.json file in your project root:

{
  "mcpServers": {
    "obsidian-lifeos": {
      "url": "http://127.0.0.1:51377/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_ACTIVATION_CODE"
      }
    }
  }
}

Available Tools

The MCP server provides the following tools for AI assistants:

File Operations

Tool NameDescription
get_file_contentGet the content of a specified file
list_filesList files in the vault
list_tagsList all tags in the vault
search_by_tagSearch content by tag
save_chatSave conversation to a note file

Task CRUD Operations

Tool NameDescription
create_taskCreate a new task with optional due date, time range, and multi-day support
list_tasksList tasks with filtering by date, status, tags, folder, and pagination
update_taskUpdate task content, date, time, or completion status
delete_taskDelete a task (including its sub-items)

Bullet (Fleeting Note) CRUD Operations

Tool NameDescription
create_bulletCreate a new bullet (fleeting note) with optional tags
list_bulletsList bullets with filtering by date, tags, folder, and pagination
update_bulletUpdate bullet content
delete_bulletDelete a bullet (including its sub-items)

Periodic Notes Query

Tool NameDescription
get_periodic_notesGet periodic notes (yearly/quarterly/monthly/weekly/daily) with optional sub-period support

Parameters:

  • year (required): Year, e.g., 2024
  • quarter: Quarter (1-4)
  • month: Month (1-12)
  • week: ISO week number (1-53)
  • day: Day of month (1-31, requires month)
  • includeQuarters: Include quarterly notes (only for yearly)
  • includeMonths: Include monthly notes (for yearly/quarterly)
  • includeWeeks: Include weekly notes (for yearly/quarterly/monthly)
  • includeDays: Include daily notes
  • includeContent: Include note content (default: true)
  • includeVirtualDocument: Include virtual document with query results

Usage Examples:

Get the 2024 yearly note
→ { year: 2024 }

Get June 2024 monthly note with all daily notes
→ { year: 2024, month: 6, includeDays: true }

Get Q2 quarterly note with monthly and weekly notes
→ { year: 2024, quarter: 2, includeMonths: true, includeWeeks: true }

Use Cases

  • Knowledge Retrieval: Let Claude search your notes directly for relevant information
  • Content Summarization: Have AI read and summarize your note content
  • Cross-tool Collaboration: In IDEs like Cursor, let AI reference your Obsidian notes while coding
  • Automated Organization: Let AI help organize and archive your notes
  • Task Management: Create, query, update, and delete tasks directly through AI assistants
  • Quick Capture: Record fleeting thoughts and ideas as bullets through AI without opening Obsidian

Important Notes

  • Make sure Obsidian is open and Deep Ask plugin is running
  • The MCP server default port is 51377
  • Authorization uses your Deep Ask activation code
  • Only supports local access (127.0.0.1) for security

🛠️ Troubleshooting

Common Issues

Q: AI does not reply or reply incorrectly? A: Check if the API Key is correct, and the network connection is normal

Q: Unable to find related note content?
A: Check if the path filtering is set correctly, and check if the Dataview plugin is working properly

Q: Plugin startup failed? A: Check if the Obsidian version meets the requirements, and the Dataview plugin is installed

Q: Activation code invalid? A: Check if the email and activation code are correct, and contact the developer for support

Log Debugging

  1. Open the developer tools (Ctrl/Cmd + Shift + I)
  2. View the error information in the Console tab
  3. Troubleshoot based on the error information

📚 Best Practices

  1. Organize knowledge structure:Establish a clear folder and tag system to facilitate AI retrieval
  2. Use context reasonably:Select related files and tags to improve answer quality
  3. Classify conversation:Create independent conversations for different topics

🚀 More Features in Development

We're constantly working on new features to enhance your experience.

New features are constantly being added! For the complete changelog, please visit Changelog.