ickyMCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@ickyMCPsearch for non-compete clauses in our partnership agreements"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
ickyMCP
RAG MCP server for semantic document search. It indexes local document collections into SQLite with sqlite-vec, then exposes MCP tools for indexing, searching, refreshing, listing, and deleting indexed documents.
Features
Semantic search over PDF, Word, PowerPoint, Excel, Markdown, and text files
Per-user database isolation with optional legacy single-database mode
Document ID filters for chat or matter-scoped retrieval
Incremental indexing based on file size and modified time
Voyage AI embeddings by default, with an offline local sentence-transformers backend
Portable SQLite storage with ignored local database artifacts
Related MCP server: MCP-RAGNAR
Setup
python -m venv .venv
.\.venv\Scripts\activate
pip install -r requirements.txtCreate local environment settings from the example:
copy .env.example .envSet ICKY_VOYAGE_API_KEY when using the default Voyage backend. For offline use, set ICKY_EMBEDDING_PROVIDER=local.
Configuration
Variable | Default | Description |
|
|
|
| unset | Required when using Voyage embeddings |
|
| Voyage embedding model |
|
| Voyage output dimension |
|
| Reserved for parallel Voyage workflows |
|
| Local sentence-transformers model |
|
| Approximate tokens per chunk |
|
| Approximate token overlap between chunks |
|
| Base directory for per-user databases |
| provider-specific | Legacy single database path |
Claude MCP Configuration
{
"mcpServers": {
"ickyMCP": {
"command": "python",
"args": ["C:\\Users\\devan\\OneDrive\\Desktop\\Projects\\ickyMCP\\run.py"],
"env": {
"ICKY_EMBEDDING_PROVIDER": "voyage",
"ICKY_VOYAGE_API_KEY": "YOUR_VOYAGE_API_KEY",
"ICKY_CHUNK_SIZE": "5000",
"ICKY_CHUNK_OVERLAP": "500"
}
}
}
}Tools
index: index a file or directory; acceptsuser_id,patterns,exclude, andforcesearch: semantic query over indexed chunks; acceptsdocument_ids,path_filter, andfile_typessimilar: find chunks similar to supplied textrefresh: re-index changed files and remove deleted files from the indexlist: list indexed documents and their IDsdelete: delete by path, document IDs, or all documentsstatus: return database, embedding, and chunking status
Verification
python -m compileall -q src run.py fast_index.py
python -c "from src.config import EMBEDDING_PROVIDER; print(EMBEDDING_PROVIDER)"The root test_*.py files are integration scripts that expect local documents under docs/ and, for the default backend, a configured Voyage API key.
Data Hygiene
Generated databases and local document folders are ignored by git: *.db, *.db.bak, *.sqlite, *.sqlite3, data/, and docs/. Keep real API keys in environment variables or .env, not in tracked files.
License
MIT
This server cannot be deployed
Maintenance
Related MCP Connectors
The Needle MCP server enables semantic search on documents stored in files like PDFs, DOCX, and XLSX by connecting AI applications to external data sources. It provides capabilities to create and manage document collections, perform natural language searches on stored content, and retrieve relevant information without requiring exact keyword matches.
Ingest, manage, and retrieve documents for RAG-powered AI applications
Vector RAG store for Word/Excel/PDF/PowerPoint. Break-even pricing, $5 per 5,700 pages.
Personal knowledge base MCP server with semantic search, auto-categorization, metadata extraction
Related MCP Servers
- AlicenseAqualityAmaintenancePrivacy-first local document search using semantic search. Runs entirely on your machine with no cloud services, supporting PDF, DOCX, TXT, and Markdown files.2293,151390MIT
- AlicenseNot gradedqualityDmaintenanceA local RAG server that enables document indexing and sentence window retrieval across multiple file formats like PDF, MD, and DOCX. It supports both local Hugging Face models and OpenAI embeddings for efficient context-aware querying through the Model Context Protocol.GPL 3.0
- AlicenseAqualityDmaintenanceA local-first document retrieval MCP server that enables AI coding tools like Codex to search private local documents via semantic search and keyword boost, supporting ingestion of PDF, DOCX, TXT, Markdown, and HTML files.7MIT
- AlicenseNot gradedqualityBmaintenanceA local-first semantic search server for documents, supporting PDFs, Office files, and text/markdown, enabling natural language search via the Model Context Protocol (MCP).1MIT