dth-houdini-mcp
MCP server exposing a live Houdini session for inspecting and debugging the DazToHue pipeline, currently providing a read-only ping tool with authenticated, loopback-only access.
README
dth-houdini-mcp
An MCP server exposing a live Houdini session for inspecting and debugging the DazToHue (DTH) pipeline.
Currently at Phase 0 — a transport spike only. One read-only tool
(houdini_ping). See PHASE0.md for what has been verified, how, and
what has not.
Layout
houdini_side/dth_listener.py runs INSIDE Houdini (started manually)
houdini_mcp/server.py stdio MCP server (launched by the MCP client)
tests/ the Phase 0 verification checks
logs/ audit log, one JSON line per executed command
Setup
python -m venv .venv
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
Run
1. Start the listener inside Houdini (Python Shell, or a shelf tool):
exec(open(r"D:\Development\dth-houdini-mcp\houdini_side\dth_listener.py").read())
start() # binds 127.0.0.1:8911
status() # inspect
stop() # shut down
2. Register the MCP server with your client:
{
"mcpServers": {
"dth-houdini": {
"command": "D:\\Development\\dth-houdini-mcp\\.venv\\Scripts\\python.exe",
"args": ["D:\\Development\\dth-houdini-mcp\\houdini_mcp\\server.py"]
}
}
}
Port is configurable on both sides via DTH_HOUDINI_MCP_PORT (default 8911).
Safety
Phase 0 is read-only: a fixed one-entry command table, loopback-only binding, and
nothing that saves or writes the .hip. Every command executed is appended to
logs/listener-<pid>.jsonl with its arguments, outcome and duration.
Requests are authenticated. start() writes a per-run token to
%LOCALAPPDATA%\dth-houdini-mcp\token; every request must carry it, and it is
checked before the command table is consulted. The token is regenerated each run
and deleted on stop().
Loopback binding is not a security boundary — a line-oriented parser is
reachable from HTTP-shaped traffic, which was measured against an earlier build
of this listener and is now blocked two ways (token, plus dropping the
connection on the first malformed line). Both are covered by
tests/test_crossprotocol.py. The full account is in
PHASE0.md §7.
License
MIT © Polynaut
Not affiliated with SideFX or with the DazToHue toolset.
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.