What embedding providers does NOUZ-MCP support?

NOUZ-MCP supports OpenAI, GigaChat, and LMStudio as embedding providers via the EMBED_PROVIDER environment variable. You specify the model name and API endpoint separately, allowing flexibility to use local or cloud-hosted embeddings.

What are the different server modes (luca, prizma, sloi)?

The MODE environment variable lets you choose between three operational modes: luca, prizma, and sloi. The README does not detail the differences; consult the project documentation or source code to understand which mode fits your use case.

How do I set up NOUZ-MCP with my Obsidian vault?

Set the OBSIDIAN_ROOT environment variable to your vault's file path, then install via `uvx nouz-mcp@2.3.0`. If you want semantic search, enable EMBED_ENABLED and configure your embedding provider (OpenAI, GigaChat, or LMStudio) with EMBED_PROVIDER, EMBED_MODEL, and EMBED_API_URL.

Can I use NOUZ-MCP without embeddings?

Yes. Embeddings are optional and controlled by the EMBED_ENABLED flag. You can run the server in DAG hierarchy and auto-classification mode without semantic search if you disable embeddings.

What is the DAG hierarchy feature?

NOUZ-MCP automatically classifies and organizes your notes into a Directed Acyclic Graph (DAG) hierarchy, creating structured relationships between documents. This enables hierarchical queries and navigation without manual folder organization.

MCP ServerSTDIOOfficialv2.3.0

Nouz MCP Server

Index and query your Obsidian vault using semantic search and auto-classified DAG hierarchies. Built for developers who want AI-powered knowledge retrieval across interconnected notes without manual tagging.

io.github.KVANTRA-dev/NOUZ-MCP

Try MCP with AI models View source

Hosted URL

Local install

Transport

STDIO

Auth

No auth required

What the Nouz MCP server does

How models use it and what it is built for.

Connect to Nouz

Local install — runs as a subprocess.

uvx nouz-mcp==2.3.0

Environment variables

Configuration this server reads at startup.

OBSIDIAN_ROOT
Path to Obsidian vault
MODE
Server mode: luca, prizma, or sloi
EMBED_ENABLED
Enable embedding-based features
EMBED_PROVIDER
Embedding provider: openai, gigachat, lmstudio
EMBED_MODEL
Embedding model name
EMBED_API_URL
Embedding API endpoint

Resources

Where to find authoritative docs and source for Nouz.

Source repository

Example prompts for Nouz

Paste any of these into Agent Studio after connecting Nouz.

Search my Obsidian vault for notes semantically similar to machine learning pipelines
Show me the DAG hierarchy of all notes tagged with architecture decisions
Query my vault using luca mode and return the top 5 most relevant documents
Explain how to configure EMBED_PROVIDER and EMBED_MODEL for custom embedding backends

Nouz MCP server — FAQ

Common questions about connecting and running Nouz.

What embedding providers does NOUZ-MCP support?
NOUZ-MCP supports OpenAI, GigaChat, and LMStudio as embedding providers via the EMBED_PROVIDER environment variable. You specify the model name and API endpoint separately, allowing flexibility to use local or cloud-hosted embeddings.
What are the different server modes (luca, prizma, sloi)?
The MODE environment variable lets you choose between three operational modes: luca, prizma, and sloi. The README does not detail the differences; consult the project documentation or source code to understand which mode fits your use case.
How do I set up NOUZ-MCP with my Obsidian vault?
Set the OBSIDIAN_ROOT environment variable to your vault's file path, then install via `uvx nouz-mcp@2.3.0`. If you want semantic search, enable EMBED_ENABLED and configure your embedding provider (OpenAI, GigaChat, or LMStudio) with EMBED_PROVIDER, EMBED_MODEL, and EMBED_API_URL.
Can I use NOUZ-MCP without embeddings?
Yes. Embeddings are optional and controlled by the EMBED_ENABLED flag. You can run the server in DAG hierarchy and auto-classification mode without semantic search if you disable embeddings.
What is the DAG hierarchy feature?
NOUZ-MCP automatically classifies and organizes your notes into a Directed Acyclic Graph (DAG) hierarchy, creating structured relationships between documents. This enables hierarchical queries and navigation without manual folder organization.

Skip the local setup — run MCP in your browser

MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 60+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.

Open Agent Studio