Practical overview
When it fits
Use it when Claude Code needs to find or retrieve material from an Obsidian vault already indexed by Smart Connections.
Who it is for
- Obsidian users who have Smart Connections generating vault embeddings and want Claude Code to retrieve notes by meaning.
- Developers configuring local MCP tools for Claude Code.
Documented uses
- Search vault notes by natural-language meaning rather than exact keywords.
- Find notes semantically related to a selected file.
- Retrieve relevant note text blocks as context for a RAG query.
Requirements & limits
The README documents local installation and configuration with Claude Code and a vault path. It describes no hosted service or account flow.
- Requires Smart Connections to have produced .smart-env/multi/*.ajson embedding files; this server does not re-index the vault.
- Uses BGE-micro-v2 query encoding and cosine similarity to rank results.
- The documented Python dependencies include MCP, sentence-transformers, NumPy below 2.0, PyTorch and Transformers.
- The server reads embeddings on startup; restart it to reload updated data.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Set up the server for Claude Code
Install the Python dependencies and configure Claude Code to connect to the server and an Obsidian vault.
Before you start
- A checked-out Smart Connections MCP Server repository
- Python environment managed with UV
- Claude Code
- An Obsidian vault with Smart Connections enabled and generated .smart-env/multi/*.ajson files
Steps
Install UV if needed, then create a virtual environment in the repository and install dependencies from requirements.txt.
Add the server entry to ~/.mcp.json, using the virtual-environment Python and server.py paths and setting OBSIDIAN_VAULT_PATH to the vault location.
Run claude mcp list to check the configured MCP connection.
Expected outcome
Claude Code is configured to access the server’s semantic search, related-note and context-block tools.
Guide limitations
- The server reads existing Smart Connections embeddings; it does not index the vault itself.
- The README specifies NumPy below 2.0 and advises using the virtual-environment Python.
- The setup instructions do not establish that every environment will connect successfully.
Compare options for Vector search
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| Smart Connections MCP Server | Search vault notes by natural-language meaning rather than exact keywords. | CLI, Self-hosted | The README documents local installation and configuration with Claude Code and a vault path. It describes no hosted service or account flow. |
| YouTube MCP Server | Search YouTube videos using a query. | Server | The project is a public GitHub repository with an MIT license. Its documented setup requires a Google API key for embeddings. |
| Milvus Skill | Create and manage Milvus collections, schemas and indexes. | CLI, Cloud, Self-hosted | The repository is public and licensed under Apache-2.0. Its documented installation method adds the skill to Claude Code. |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 18 October 2026
- Original website title
- dan6684/smart-connections-mcp · Official homepage title · retained in its source language

