klayout-mcp
Read-only MCP server for KLayout that opens GDS/OAS layouts, inspects geometry and hierarchy, renders deterministic PNGs, and runs batch DRC with structured JSON output.
README
klayout-mcp
Read-only MCP server for KLayout. It opens GDS/OAS layouts, inspects geometry and hierarchy, renders deterministic PNGs, and runs batch DRC with structured JSON output.
Install
Fastest one-off run:
uvx klayout-mcp@latest
Install once:
uv tool install klayout-mcp
Traditional Python install:
python -m pip install klayout-mcp
For batch DRC, the klayout executable must be on PATH or exposed through KLAYOUT_BIN.
export KLAYOUT_BIN=/Applications/klayout.app/Contents/MacOS/klayout
Quick Start
klayout-mcp is a stdio MCP server. The default launch shape is:
command: uvx
args: klayout-mcp@latest
If your client cannot use uvx, install the tool once and launch klayout-mcp directly.
Typical workflow:
open_layoutlist_layerslist_cellsordescribe_cellquery_regionmeasure_geometryanalyze_waveguiderender_viewrun_drc_scriptextract_markersclose_session
Client Setup
More copy-paste examples live in examples/mcp/README.md.
Codex
[mcp_servers.klayout]
command = "uvx"
args = ["klayout-mcp@latest"]
[mcp_servers.klayout.env]
KLAYOUT_BIN = "/Applications/klayout.app/Contents/MacOS/klayout"
Or add it with the CLI:
codex mcp add klayout \
--env KLAYOUT_BIN=/Applications/klayout.app/Contents/MacOS/klayout \
-- uvx klayout-mcp@latest
Claude Code
{
"mcpServers": {
"klayout": {
"command": "uvx",
"args": ["klayout-mcp@latest"],
"env": {
"KLAYOUT_BIN": "/Applications/klayout.app/Contents/MacOS/klayout"
}
}
}
}
Or add it with the CLI:
claude mcp add --transport stdio klayout \
--scope project \
--env KLAYOUT_BIN=/Applications/klayout.app/Contents/MacOS/klayout \
-- uvx klayout-mcp@latest
Cursor
Cursor is file-config based:
{
"mcpServers": {
"klayout": {
"command": "uvx",
"args": ["klayout-mcp@latest"],
"env": {
"KLAYOUT_BIN": "/Applications/klayout.app/Contents/MacOS/klayout"
}
}
}
}
OpenCode
OpenCode is file-config based:
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"klayout": {
"type": "local",
"command": ["uvx", "klayout-mcp@latest"],
"enabled": true,
"environment": {
"KLAYOUT_BIN": "/Applications/klayout.app/Contents/MacOS/klayout"
}
}
}
}
Other MCP Hosts
Use:
command:uvxargs:["klayout-mcp@latest"]env:KLAYOUT_BINwhen needed
Tools
- Session:
open_layout,close_session - Structure:
list_layers,list_cells,describe_cell - Geometry:
query_region,measure_geometry,analyze_waveguide - View:
set_view,render_view - DRC:
run_drc_script,extract_markers
Configuration
| Variable | Purpose | Default |
|---|---|---|
KLAYOUT_MCP_ARTIFACT_ROOT |
Root directory for runtime artifacts | <repo>/.artifacts |
KLAYOUT_MCP_SESSION_TTL_SECONDS |
Session inactivity timeout | 3600 |
KLAYOUT_BIN |
KLayout batch executable for DRC | klayout |
Behavior:
- any absolute local path can be used for layouts and DRC scripts
- artifact paths returned by tools are absolute
- sessions expire lazily after inactivity
close_sessionremoves the session artifact directory
Artifacts And Errors
Artifacts are stored under .artifacts/sessions/<session_id>/ and include renders, DRC reports, marker crops, and logs.
Tool failures return structured JSON such as:
{
"code": "FILE_NOT_FOUND",
"message": "Layout file does not exist",
"details": {
"path": "/abs/path/to/missing.gds"
}
}
Common codes: FILE_NOT_FOUND, SESSION_NOT_FOUND, INVALID_BOX, INVALID_LAYER, INVALID_TARGET, DRC_RUN_FAILED.
Development
For a source checkout:
uv sync --extra dev
uv run klayout-mcp
If an MCP client needs to launch the checkout directly:
command: uv
args: --directory /abs/path/to/klayout-mcp run klayout-mcp
Contributor workflow is in CONTRIBUTING.md. Release notes are in CHANGELOG.md.
Releases
Normal releases are automated with release-please.
- merge Conventional Commits to
main release-pleaseopens or updates a release PR- merging that release PR updates
pyproject.toml,CHANGELOG.md, creates the tag and GitHub release, and publishes to PyPI
To let the release PR run normal CI under branch protection, configure a repository secret named RELEASE_PLEASE_TOKEN with a GitHub token that can open pull requests. Without it, release-please falls back to github.token, which may not trigger PR workflows.
The manual release.yml workflow remains available for TestPyPI validation and recovery publishing.
Reference
- docs/specs/2026-03-05-klayout-observer-mcp-contract.md
- docs/plans/2026-03-05-klayout-observer-mcp-design.md
- docs/plans/2026-03-05-klayout-observer-mcp-implementation-plan.md
Documentation Site
The repository is configured for Read the Docs with MkDocs.
Build the docs locally:
uv run --extra docs mkdocs build --strict
Serve the docs locally:
uv run --extra docs mkdocs serve
After importing the GitHub repository into Read the Docs, it can build directly from the default branch using .readthedocs.yaml.
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.