MCP Tools
add_note
Create markdown notes with automatic embedding
Create new notes in your knowledge base with automatic chunking and embedding generation.
Parameters
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
title | string | REQUIRED | - | Note title (1-500 chars) |
content | string | REQUIRED | - | Markdown content (1-100000 chars) |
tags | string[] | OPTIONAL | [] | Tags for organization |
extractMemories | boolean | OPTIONAL | false | Extract entities and facts as memories (REQUIRES ENABLE_MEMORY_EXTRACTION=true and ANTHROPIC_API_KEY) |
Example Request
{
"title": "Project Alpha Requirements",
"content": "# Project Alpha\n\n## Goals\n\n- Improve search accuracy by 20%\n- Reduce latency to under 100ms\n\n## Technical Requirements\n\n- Use hybrid search with RRF\n- Implement caching layer\n- Add monitoring dashboards",
"tags": ["project-alpha", "requirements", "technical"]
}Response
{
"documentId": "550e8400-e29b-41d4-a716-446655440000",
"title": "Project Alpha Requirements",
"chunkCount": 1,
"tags": ["project-alpha", "requirements", "technical"],
"createdAt": "2024-11-01T10:00:00Z"
}Processing Pipeline
When you add a note:
- Document Created: Stored in
documentstable - Content Chunked: Split into ~512 token chunks with overlap
- Embeddings Generated: Each chunk embedded via OpenAI/Ollama
- Search Index Updated: Full-text search vector created
Use Cases
Knowledge Capture
Save insights from conversations:
{
"title": "Customer Feedback - October 2024",
"content": "Key themes from customer calls:\n\n1. Users want faster search\n2. Mobile app requested\n3. Integration with Slack needed",
"tags": ["customer-feedback", "product"]
}Meeting Notes
Capture meeting outcomes:
{
"title": "Sprint Planning - Week 45",
"content": "## Attendees\n- Team leads\n\n## Decisions\n- Prioritize search improvements\n- Delay mobile app to Q1",
"tags": ["meetings", "sprint-planning"]
}Research Notes
Store research findings:
{
"title": "Hybrid Search Research",
"content": "## Papers Reviewed\n\n1. RRF paper - key insights...\n2. pgvector benchmarks...",
"tags": ["research", "search", "technical"]
}Error Responses
Content too long
{
"error": "Validation error",
"message": "Content exceeds maximum length of 100000 characters"
}Embedding failed
{
"error": "Embedding failed",
"message": "OpenAI API rate limit exceeded"
}