Kuksa MCP Server
A Model Context Protocol (MCP) server that connects LLMs to vehicle data via the Eclipse Kuksa Databroker, enabling AI assistants to read and write vehicle signals using the standardized COVESA Vehicle Signal Specification.
README
Kuksa MCP Server
A Model Context Protocol (MCP) server that connects LLMs to vehicle data via the Eclipse Kuksa Databroker. Enables AI assistants to read and write vehicle signals using the standardized COVESA Vehicle Signal Specification (VSS).
Features
- Read signals — Get current or target values of any VSS signal
- Batch reads — Fetch multiple signals in a single call
- Write actuators — Set actuator targets or publish sensor readings
- Browse the VSS tree — List or count signals under any branch, with substring filtering
- Inspect types — Query data types for any signal path
- Server introspection — Name, version, signal count at startup
- Per-request logging — Every tool call logged with params, timing, and result summary
- Dual transport — stdio (default) and SSE (HTTP)
- OpenCode ready — Pre-configured
opencode.jsoncfor local and remote variants
Quick Start
1. Start a Kuksa Databroker
docker run -d --rm -p 55555:55555 --name kuksa-databroker \
ghcr.io/eclipse-kuksa/kuksa-databroker:main --insecure
2. Install & Run
pip install kuksa-mcp-server
kuksa-mcp
Or from source:
git clone https://github.com/mikehaller/kuksa-mcp-server.git
cd kuksa-mcp-server
pip install -e ".[dev]"
kuksa-mcp
3. Configure your MCP client
<details> <summary>Claude Desktop / Cursor (<code>claude_desktop_config.json</code>)</summary>
{
"mcpServers": {
"kuksa": {
"command": "kuksa-mcp"
}
}
}
</details>
<details> <summary>OpenCode — stdio variant (<code>opencode.jsonc</code> in project root)</summary>
{
"mcpServers": {
"kuksa-local": {
"type": "local",
"command": ["kuksa-mcp", "--log-file", "./kuksa-mcp.log"],
"enabled": true,
"timeout": 10000
}
}
}
</details>
<details> <summary>OpenCode — SSE variant (<code>opencode.jsonc</code> in project root)</summary>
{
"mcpServers": {
"kuksa-remote": {
"type": "remote",
"url": "http://127.0.0.1:8765/sse",
"enabled": false,
"timeout": 10000
}
}
}
</details>
OpenCode MCP Server Variants
The project ships with two entries in opencode.jsonc:
| Variant | Transport | Default | Use case |
|---|---|---|---|
kuksa-local |
stdio (subprocess) | enabled | Local dev, single-user, auto-managed |
kuksa-remote |
SSE (HTTP) | disabled | Multi-user, containerized, remote databroker |
Start the remote variant manually:
kuksa-mcp --transport sse --host 0.0.0.0 --port 8765
CLI Reference
kuksa-mcp [options]
Options:
--transport <stdio|sse> Transport protocol (default: stdio)
--host <ip> Bind address for SSE (default: 127.0.0.1)
--port <port> Port for SSE (default: 8765)
--kuksa-host <host> Databroker host (overrides KUKSA_HOST env)
--kuksa-port <port> Databroker port (overrides KUKSA_PORT env)
--log-file <path> File path for request logging (also written to stderr)
Environment Variables
| Variable | Default | Description |
|---|---|---|
KUKSA_HOST |
127.0.0.1 |
Kuksa Databroker host |
KUKSA_PORT |
55555 |
Kuksa Databroker gRPC port |
KUKSA_TOKEN |
(none) | JWT token for authorization |
Tools
| Tool | API | Description |
|---|---|---|
get_signal(path) |
Kuksa V2 | Get current value of one VSS signal |
get_signals(paths) |
Kuksa V2 | Get current values of multiple signals |
set_signal(path, value, datatype) |
Kuksa V2 | Set an actuator target value |
publish_value(path, value, datatype) |
Kuksa V2 | Publish a sensor reading (provider role) |
list_signals(branch, query) |
V1 (deprecated) | List signals and metadata under a branch |
count_signals(branch, query) |
V1 (deprecated) | Count signals matching a filter |
get_target_values(paths) |
V1 (deprecated) | Read actuator target/desired values |
get_value_types(paths) |
V1 (deprecated) | Get data types for one or more signals |
server_info() |
V1 (deprecated) | Get databroker name, version, address |
Resources
| URI | Description |
|---|---|
kuksa://info |
Server and tool reference |
kuksa://signals/{path} |
Signal info (redirects to tool) |
kuksa://branches/{path} |
Branch listing (redirects to tool) |
Startup Output
On startup the server prints a banner with version, databroker info, and catalog stats:
======================================================
Kuksa Databroker MCP Server
Version : 0.1.0
Built : 2026-06-24
======================================================
Backend : databroker
Address : 127.0.0.1:55555
Version : 0.7.0-dev.0
Status : connected
Signals : 1263 total
Types : ACTUATOR=643, ATTRIBUTE=130, SENSOR=490
======================================================
Every tool call is logged:
09:45:12 [INFO] Processing request of type ListToolsRequest
09:45:12 [INFO] Processing request of type CallToolRequest
09:45:12 [INFO] get_signal(args={'path': 'Vehicle.Speed'}) -> ok [0.023s]
09:45:15 [INFO] list_signals(args={'branch': 'Vehicle.Cabin', 'query': ''}) -> 490 results [0.045s]
Development
pip install -e ".[dev]"
pytest
kuksa-mcp --transport sse
python examples/local_test.py
Docker
The server is published as a Docker image on GitHub Container Registry:
docker pull ghcr.io/mikehaller/kuksa-mcp-server:latest
# Run with SSE transport
docker run --rm -p 8765:8765 \
-e KUKSA_HOST=host.docker.internal \
ghcr.io/mikehaller/kuksa-mcp-server:latest \
--transport sse --host 0.0.0.0 --port 8765
The image uses a non-root kuksa user and is built from python:3.12-slim (~200 MB).
License
Apache 2.0 — see LICENSE.
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.