jlcpcb-search-mcp
Enables AI assistants to search JLCPCB's 450K+ electronic component catalog with real-time stock levels and pricing, supporting keyword search, filtering, and detailed component information.
README
JLCPCB Search MCP Server
A Model Context Protocol (MCP) server that provides AI assistants with powerful search capabilities for JLCPCB's electronic component catalog. Search 450,000+ components with real-time stock levels and pricing.
Built on top of yaqwsx/jlcparts - huge thanks to Jan Mrazek for doing the hard work of downloading JLCPCB XLS sheets, converting them to structured JSON, and hosting them for the community!
Features
- Fast Local Search: SQLite database with 450K+ components for instant searches
- Live Data: Real-time stock levels and pricing from JLCPCB API
- Smart Filtering: Filter by package type, category, Basic/Extended parts, minimum stock
- Detailed Information: Full specifications, datasheets, pricing tiers, and images
- Easy Integration: Works with Claude Desktop, VSCode, and any MCP-compatible client
Example Usage
Ask Claude natural questions about components:
You: "Can you find a good 10uF SMD capacitor that can support 10V or less?"
Claude's Response:
Based on my search, here are some good 10uF SMD capacitors that can support 10V or less:
Recommended Options (Aluminum Electrolytic SMD):
Best Choice - C4747968 (RST10UF16V013)
- Rating: 10uF, 16V
- Package: SMD, D4xL5.4mm
- Stock: 76,080 units (excellent availability)
- Price: $0.0248 (20+), $0.0170 (600+)
- Link: https://jlcpcb.com/partdetail/C4747968
Alternative - C970654 (RVT1C100M0405)
- Rating: 10uF, 16V
- Package: SMD, D4xL5.4mm
- Stock: 7,360 units
- Price: $0.0293 (20+), $0.0202 (600+)
- Link: https://jlcpcb.com/partdetail/C970654
All of these are aluminum electrolytic SMD capacitors rated for 16V, which gives you plenty of margin for a 10V application. The C4747968 is my top recommendation due to its excellent stock levels and lower price.
Note: These are all Extended parts (not Basic), so they will incur additional assembly fees if you're using JLCPCB assembly.
Tools
The MCP server exposes four tools to your AI assistant:
| Tool | What it does |
|---|---|
search_components |
Keyword search across the catalog. Optional filters: category, package, basic_only, min_stock, max_results. |
get_component_details |
Full live details for an LCSC part — current stock, pricing tiers, full parameter list, datasheet, and images. |
database_status |
Reports the local database location, size, and component count. |
refresh_database |
Re-downloads the catalog snapshot from inside the chat — no CLI needed. |
Most of the time you don't call these directly; you ask Claude in natural language ("find me a 10k 0805 resistor with at least 5k in stock") and it picks the right tool.
Data Freshness
This server uses a hybrid approach — a local catalog snapshot for fast search, plus live API calls for stock and pricing.
| Data | Source | Freshness |
|---|---|---|
| Component catalog (descriptions, packages, attributes, categories) | Local SQLite, built from yaqwsx/jlcparts | Snapshot from your last refresh |
| Stock levels | Live JLCPCB API (wmsc.lcsc.com) |
Real-time, per query |
| Pricing tiers | Live JLCPCB API | Real-time, per query |
| Datasheet URL | Live JLCPCB API | Real-time, per query |
The local catalog does not auto-update. Refresh it periodically to pick up newly-added parts, removed parts, and metadata changes:
jlcpcb-mcp-setup --refresh-db
For most workflows, refreshing weekly or monthly is sufficient. Stock and price decisions are always made against live data, so you don't need a fresh catalog to trust the numbers on parts you already know.
Installation
From PyPI
# Install for current user (recommended for Claude Desktop)
pip install --user jlcpcb-search-mcp
# Or use pipx for isolated installation (best for CLI tools)
pipx install jlcpcb-search-mcp
# Or install in a virtual environment
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install jlcpcb-search-mcp
# Or install with uv in a project (uv is my fav!)
uv add jlcpcb-search-mcp
From Source
git clone https://github.com/peterb154/jlcpcb-search-mcp.git
cd jlcpcb-search-mcp
pip install -e .
Important: When using Claude Code, the MCP server must be accessible from your project's environment:
- For uv projects: Install with
uv add jlcpcb-search-mcpand useuv run claudecommands - For system-wide use: Install with
pip install --user jlcpcb-search-mcporpipx install jlcpcb-search-mcp - For venv projects: Install in your venv and activate it before starting your editor
Quick Start
Claude Code Configuration
Option 1: Automatic Setup (Recommended)
Use the setup utility to automatically configure the MCP server:
# For current workspace (creates .mcp.json)
# Uses shared database in user's application data directory - RECOMMENDED
jlcpcb-mcp-setup --workspace
# Or if using uv project
uv run jlcpcb-mcp-setup --workspace
# For development of THIS package only (uses ./data/ directory)
# This creates a separate 900MB+ database in your project - avoid unless needed
jlcpcb-mcp-setup --workspace --dev
# Globally for Claude Desktop (also uses shared database)
jlcpcb-mcp-setup --global
After running the setup:
-
Reload your editor window (Cmd+Shift+P → Developer: Reload Window)
-
Check that the MCP server is connected:
# If using uv project: uv run claude mcp listshould return something like the following - notice that the mcp server is connected:
Checking MCP server health... jlcpcb-search: jlcpcb-mcp - ✓ Connected# Otherwise if not using uv: claude mcp list -
You should see:
jlcpcb-search: jlcpcb-mcp - ✓ Connected -
Start searching for components!
Option 2: Manual Configuration
Create .mcp.json in your workspace root:
{
"mcpServers": {
"jlcpcb-search": {
"command": "jlcpcb-mcp"
}
}
}
Or for global configuration in Claude Desktop:
macOS: ~/Library/Application Support/Claude/mcp.json
Windows: %APPDATA%\Claude\mcp.json
Linux: ~/.config/claude/mcp.json
Note for uv projects: The same simple config works! When you run uv run claude, it automatically uses the environment where jlcpcb-mcp is installed. No special configuration needed.
First Run & Database Setup
The database will be automatically built on your first component search. However, you can pre-download it to avoid waiting:
# Pre-download the database (recommended)
jlcpcb-mcp-setup --refresh-db
# Or with uv
uv run jlcpcb-mcp-setup --refresh-db
What to expect:
- Download: ~50MB compressed (manifest + attributes lookup table + ~1,294 per-subcategory shards)
- Final size: ~1.8GB SQLite database uncompressed (~576k components, 1,289 subcategories)
- Time: ~4 minutes on fast internet connection (500Mbps), up to 10 minutes on slower connections
- Progress: Real-time updates showing download progress and subcategory processing
If you don't pre-download, your first Claude query will trigger the download automatically, so expect a 3-10 minute wait.
Database Location
The database is shared across all your projects by default (recommended):
- macOS:
~/Library/Application Support/jlcpcb-mcp/components.sqlite - Linux:
~/.local/share/jlcpcb-mcp/components.sqlite - Windows:
%LOCALAPPDATA%\jlcpcb-mcp\components.sqlite
Why shared?
- One ~1.8GB database serves all your projects
- Saves disk space
- Faster setup for new projects
- Easier to keep updated
Dev mode (--dev): Only use this if you're developing the jlcpcb-mcp package itself. It creates a separate database in ./data/ which is useful for testing but wastes space for normal usage.
Custom location (JLCPCB_DATABASE_PATH)
If you want the database to live somewhere specific — e.g. an external SSD, a project-local directory, or a path you can .gitignore — set the JLCPCB_DATABASE_PATH environment variable. The value is the full path to the SQLite file (not just the directory):
// .mcp.json (or your client's MCP config)
{
"mcpServers": {
"jlcpcb-search": {
"command": "jlcpcb-mcp",
"env": {
"JLCPCB_DATABASE_PATH": "/path/to/your/components.sqlite"
}
}
}
}
To pre-download into that location, export the same variable before running setup:
JLCPCB_DATABASE_PATH=/path/to/your/components.sqlite \
jlcpcb-mcp-setup --refresh-db
Or skip pre-downloading and let the server build it on the first query.
JLCPCB_DATABASE_PATH takes precedence over both the default platform location and --dev mode.
License
MIT License - see LICENSE file for details
Acknowledgments
- yaqwsx/jlcparts by Jan Mrazek - Component data source and inspiration
- JLC Parts Web Search - If you want a great search UI without AI
- FastMCP - MCP server framework
- JLCPCB - Component data and API
Support
- Issues: GitHub Issues
- Documentation: docs/
- MCP Protocol: Model Context Protocol
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.