Skip to main content

Quick Start Guide

💡TL;DR

Start using Notemd in 5 minutes. Three steps: (1) Add an LLM API key (or use local Ollama), (2) Open a note and run "Process file (add links)", (3) Explore one-click workflows in the sidebar. Supports 30+ providers. No coding required. Works offline with Ollama.

Step 1: Configure Your First LLM Provider (2 min)​

Option A: Cloud Provider (OpenAI, Anthropic, etc.)​

  1. Open Settings → Notemd
  2. Under LLM Provider Configuration, select a provider from the dropdown
  3. Enter your API Key
  4. (Optional) Click "Get Model List" to see available models
  5. Select a model for each task type

Popular choices:

  • OpenAI: gpt-4o-mini (fast, affordable)
  • Anthropic: claude-sonnet-4-6 (high quality)
  • DeepSeek: deepseek-chat (very affordable)

Option B: Local Ollama (No API Key Needed)​

  1. Install Ollama
  2. Download a model: ollama pull llama3.1
  3. In Notemd settings, select "Ollama" as provider
  4. Model name: llama3.1 (or any model you downloaded)

Step 2: Try Your First Task (1 min)​

  1. Open any note with meaningful content
  2. Right-click in the editor
  3. Select "Process file (add links)"
  4. Wait a few seconds
  5. Key concepts are now [[wiki-linked]]!

Example:

Before:

Machine learning models use neural networks to process data.

After:

[[Machine learning]] models use [[neural networks]] to process data.

Step 3: Explore Features (2 min)​

Extract Concept Notes​

  1. Right-click → "Extract concepts from current file"
  2. Notemd creates separate files for each concept
  3. Check your configured concept folder (default: concepts/)
  4. Each concept note has a backlink to your original note

Research & Summarize​

  1. Highlight a topic in your note
  2. Right-click → "Research and summarize"
  3. Notemd queries web search (Tavily or DuckDuckGo)
  4. Summary appends to your note with sources

One-Click Workflows​

  1. Check the left sidebar for workflow buttons
  2. Click "One-Click Extract" on any note
  3. It runs: add links → extract concepts → research → generate diagram
  4. All in one click!

Common First Tasks​

Task 1: Process a Paper/Article​

1. Paste article content into a note
2. Run "Process file (add links)"
3. Run "Extract concepts"
4. Run "Research and summarize" on unfamiliar terms
5. Run "Summarize as Mermaid diagram" for visual structure

Task 2: Translate Notes​

1. Highlight text to translate
2. Right-click → "Translate to target language"
3. Set target language in settings first

Task 3: Create a Knowledge Map​

1. Process 5-10 related notes with "add links"
2. Run "Extract concepts" on each
3. Open Graph view to see connections
4. Your knowledge graph is forming!

Customization Tips​

Per-Task Model Selection​

You can use different models for different tasks:

TaskRecommended ModelWhy
Add Linksgpt-4o-mini, deepseek-chatFast, many calls
Extract Conceptsclaude-sonnet-4-6, gpt-4oQuality extraction
Researchperplexity, gpt-4oWeb-grounded
Translationgemini-flash, gpt-4oMultilingual
Diagramsclaude-opus, local modelComplex reasoning

Configure in: Settings → Notemd → Task-Specific Models

Custom Workflows​

Create your own button:

  1. Settings → Notemd → Workflows
  2. Click "Add Workflow"
  3. Enter DSL: process-current-add-links>extract-concepts-current>research-and-summarize
  4. Name it: "My Research Flow"
  5. New button appears in sidebar

Troubleshooting​

"API Key Invalid"​

  • Check if key is correct (no extra spaces)
  • Verify account has credits
  • Test with "Connection Test" button in settings

"Model Not Found"​

  • Click "Get Model List" to see available models
  • Some providers require specific model names (e.g., gpt-4o not gpt-4)
  • Ensure note has enough content (>50 words)
  • Try a more powerful model
  • Check if LLM call succeeded (see diagnostics panel)

Slow Performance​

  • Use faster models (gpt-4o-mini instead of gpt-4o)
  • Reduce chunk size in settings
  • Enable parallel processing for batch tasks

Next Steps​

Video Tutorial​

Watch the 3-minute quickstart video (Chinese): 【全开源免费！NotEMD 三分钟速通】

Getting Help​