Three plugins turn Obsidian into an AI-powered vault: Copilot for chatting with your notes, Smart Connections for surfacing related ideas automatically, and Text Generator for inline writing help. Setup takes about 15 minutes and costs $3-10/month in API usage, or nothing at all if you run a local model instead.
Quick Verdict
- Start with just Copilot — it’s the highest-value plugin and the others are optional add-ons.
- Budget $3-10/month in API costs for moderate use, or use Ollama for a free local model.
- Setup takes about 15 minutes total across all three plugins.

The Three Plugins at a Glance
| Plugin | What It Does | Setup Time |
|---|---|---|
| Copilot | Chat sidebar for asking questions about your vault | 10 minutes |
| Smart Connections | Surfaces semantically related notes automatically | 5 minutes |
| Text Generator | Inline AI writing assistance, triggered by shortcut | 5 minutes |
How I Evaluated This
I judged these plugins on how much AI capability they add relative to the setup effort, since Obsidian users tend to value simplicity and data ownership over adding complexity. The API cost estimates below are reasonable ballparks based on current OpenAI and Anthropic pricing, not fixed guaranteed rates — actual cost depends on your usage.
What You’ll Need
- Obsidian installed (free, any platform)
- An API key from OpenAI (ChatGPT) or Anthropic (Claude)… or use a free local model
- 15-20 minutes for initial setup
- A vault with some notes in it (the AI is more useful when it has content to work with)
The Three Best AI Plugins for Obsidian
1. Copilot for Obsidian — Best Overall AI Plugin
What it does: Adds a chat sidebar to Obsidian where you can ask questions about your notes, generate content, and get AI assistance. Think of it as having ChatGPT or Claude built into your note-taking app, but with access to your entire vault.
Setup (10 minutes):
- Open Obsidian Settings → Community Plugins → Browse
- Search “Copilot”
- Install and enable “Copilot”
- Go to Copilot settings
- Add your API key (OpenAI for ChatGPT or Anthropic for Claude)
- Choose your default model (GPT-4o, Claude Sonnet, or Claude Opus)
What you can do:
- Chat with your vault: “What are my notes about weekly reviews?” Copilot searches your vault and answers based on your actual notes.
- Generate content: “Write a summary of everything I’ve written about AI agents.” The AI pulls from your notes to create the summary.
- Ask questions: “Based on my project notes, what’s the next step for the client engagement?” Context-aware answers.
- Summarize notes: Select a long note and ask Copilot to summarize it.
Why it’s #1: Copilot is the most polished AI chat experience in Obsidian. The vault-aware context means answers reference your actual notes, not generic internet knowledge. It supports multiple AI models so you can use whichever you prefer.
Cost: The plugin is free. You pay for the AI API usage. ChatGPT API costs about $0.01-0.03 per query for GPT-4o. Claude API is similar. Monthly cost for moderate use: $2-10.
2. Smart Connections — Best for Discovering Links
What it does: Analyzes your vault and surfaces notes that are semantically related to whatever you’re currently writing. It’s like having an AI that reads your entire vault and says “hey, this note from three months ago is related to what you’re working on now.”
Setup (5 minutes):
- Install “Smart Connections” from Community Plugins
- Enable the plugin
- Let it index your vault (takes a few minutes depending on vault size)
- Open the Smart Connections panel from the right sidebar
What you can do:
- See related notes in real-time: As you write or view a note, the sidebar shows semantically similar notes from your vault
- Discover forgotten connections: Find notes you wrote months ago that relate to your current work
- Chat with your notes: Ask questions and get answers grounded in your vault content
- Build knowledge maps: Understand how ideas across your vault connect
Why it’s valuable: This is the plugin that makes Obsidian’s graph view smarter. Instead of just showing which notes link to each other (which requires manual linking), Smart Connections shows which notes are conceptually related even without explicit links. The unexpected connections are often the most valuable.
Cost: Free plugin. Uses AI embedding models for semantic analysis. Minimal API cost (a few cents to index a large vault, then pennies per day for ongoing use).
3. Text Generator — Best for Writing Assistance
What it does: Adds AI text generation directly into your notes. Place your cursor anywhere, trigger the plugin, and it generates text based on the context of your note. Autocomplete on steroids.
Setup (5 minutes):
- Install “Text Generator” from Community Plugins
- Add your OpenAI or Anthropic API key in settings
- Configure your preferred model and generation settings
- Set a keyboard shortcut for quick triggering (I use Cmd+J)
What you can do:
- Continue writing: Place cursor at the end of a paragraph, trigger Text Generator, and it continues your thought
- Generate from templates: Create a template prompt and generate content from it
- Summarize sections: Select text and generate a summary
- Brainstorm: Write a question in your note and generate multiple answers
- Translate: Select text and generate a translation
Why it’s useful: Text Generator is the least disruptive AI plugin. It works inline, inside your notes, without opening a sidebar or chat window. You’re writing, you need AI help, you press a shortcut, the text appears. Minimal workflow interruption.
Cost: Free plugin. API costs same as Copilot ($2-10/month for moderate use).
The Free Alternative: Local AI Models
If you don’t want to pay for API access, you can run AI models locally on your computer using Ollama. Several Obsidian plugins support local models.
Setup:
- Install Ollama (free, ollama.ai)
- Download a model:
ollama pull llama3(or any supported model) - Configure your Obsidian AI plugin to use the local Ollama endpoint instead of OpenAI/Anthropic
Pros: Completely free. Your data never leaves your computer. No API key needed.
Cons: Requires decent hardware (8GB+ RAM minimum, 16GB recommended). Model quality is below GPT-4o or Claude Opus. Slower responses on older machines.
For Obsidian users who chose the platform for privacy reasons, local AI preserves that philosophy completely.
My Recommended Setup
If I were setting up Obsidian with AI from scratch:
- Install Copilot as the primary AI interface (for chatting with your vault)
- Install Smart Connections for discovering related notes
- Use Claude as the AI model (better reasoning and writing quality)
- Set a daily budget alert at $5/month for API costs (you probably won’t hit it)
Total setup time: 15 minutes. Monthly cost: $3-10 in API usage.
This gives you:
- Chat with your entire vault (Copilot)
- Automatic discovery of related notes (Smart Connections)
- High-quality writing assistance using Claude
- Full data ownership (your notes stay local, only queries go to the API)
Obsidian AI vs. Notion AI
The question people ask: “Should I just use Notion with built-in AI instead?”
Choose Obsidian + AI plugins if:
- You want your data stored locally (not on company servers)
- You want to choose your AI model (Claude, GPT-4, local models)
- You want to control costs (pay-per-query vs flat subscription)
- You value the plugin ecosystem and customization
- Offline access matters to you
Choose Notion AI if:
- You want AI that works without any setup
- You need team collaboration
- You want databases, projects, and tasks alongside notes
- You prefer a monthly subscription over pay-per-query
- You don’t want to manage API keys
For most people, Notion’s built-in AI is simpler. For people who chose Obsidian deliberately (privacy, speed, data ownership, customization), the plugin approach preserves those values while adding AI capability.
For a deeper comparison, see Notion vs Obsidian (2026).
Step-by-Step: Your First AI-Powered Note
Let’s put it all together with a practical example.
Scenario: You just attended a lecture and want to process your notes.
- Create a new note in Obsidian with your raw lecture notes
- Open the Copilot sidebar and ask: “Summarize the key points from this note”
- Review the summary at the top of your note
- Ask follow-up questions: “What concepts in this note connect to my other notes about {topic}?”
- Check Smart Connections in the sidebar… it shows related notes you’ve written before
- Link the related notes by clicking through and adding backlinks
- Use Text Generator to expand on any point: place cursor after a key concept, press Cmd+J, and let AI elaborate
In 10 minutes, you’ve gone from raw notes to a summarized, connected, AI-enhanced knowledge entry. That same process without AI takes 30+ minutes of manual review and linking.
Common Issues and Fixes
“The AI doesn’t know about my notes.”
Make sure you’re using Copilot or Smart Connections (which index your vault), not just Text Generator (which only sees the current note). Vault-wide context requires indexing.
“API costs are adding up.”
Switch to a cheaper model (GPT-4o Mini or Claude Haiku) for routine queries. Save the expensive models (Opus, GPT-4o) for complex analysis. Or use Ollama for free local processing.
“Responses are slow.”
Local models via Ollama can be slow on older hardware. Cloud API responses are typically 2-5 seconds. If cloud responses are slow, check your internet connection or try a different model.
“I don’t want my notes sent to OpenAI/Anthropic.”
Use Ollama with local models. Your notes never leave your computer. Quality is lower but privacy is absolute.
FAQ
How much does AI in Obsidian actually cost per month?
For moderate use (10-20 queries per day): $3-10/month in API costs. For light use (a few queries per day): under $2/month. For free: use Ollama with local models.
Do I need all three plugins?
No. Start with just Copilot. Add Smart Connections after a week if you want automatic related note discovery. Add Text Generator if you want inline AI writing assistance. Each one stands alone.
Will these plugins slow down Obsidian?
Minimally. Smart Connections does an initial vault indexing that takes a few minutes on large vaults, but after that the impact is negligible. Obsidian remains fast.
Can I use these with Obsidian Sync?
Yes. The plugins store their data locally alongside your vault. Obsidian Sync handles the notes. The plugin configurations stay on each device (you’ll need to set up API keys on each device separately).
For how Obsidian compares to other second brain tools, see Best Second Brain Apps (2026).
Next Step
Recommended for you
Want the full system? 25X is the flagship productivity system we teach.
Explore 25X →