Home
Last Updated: December 6, 2025 - Production/Stable v2.6.4
🤖 Can't find what you're looking for? Try our AI-Powered Wiki Search - Ask questions in natural language and get instant answers with source attribution. Example: "How do I prevent SQL injection attacks?"
The SQLite MCP Server transforms SQLite into a powerful, AI-ready database engine. It combines standard relational operations with advanced analytics, text and vector search, geospatial capabilities, and intelligent workflow automation. By layering business intelligence tools, semantic resources, and guided prompts on top of SQLite, it enables both developers and AI assistants to interact with data more naturally and effectively.
This makes it not just a database interface, but a workflow-aware assistant for developers and AI systems.
🎛️ Tool Filtering - Selectively enable/disable tools via SQLITE_MCP_TOOL_FILTER environment variable:
- Address MCP client tool limits (Windsurf's 100-tool limit, Cursor stability)
- Reduce token overhead by exposing only needed tools
- Group-based filtering (
-vector,-stats) and individual tool control (+vacuum_database)
See Tool Filtering for complete documentation and Changelog for version history.
- Quick-Start - Installation, configuration, and first steps
- Core-Database-Tools - Essential CRUD operations (15 tools)
- JSON-Helper-Tools - v2.6.0+: Complete JSON manipulation suite (6 tools)
- Security-and-Data-Integrity - Parameter binding, JSONB, transactions
- Advanced-Text-Processing - Regex, fuzzy matching, similarity (8 tools)
- Statistical-Analysis - Descriptive stats, percentiles, time series (8 tools)
- Full-Text-Search - FTS5 implementation with BM25 ranking (3 tools)
- Backup-and-Restore - Enterprise-grade backup operations (5 tools)
- PRAGMA-Operations - Database configuration and optimization (5 tools)
- Virtual-Tables - R-Tree spatial indexing, CSV import/export (6 tools)
- Enhanced-Virtual-Tables - Smart CSV/JSON import with schema inference (4 tools)
- Semantic-Vector-Search - AI-native embeddings and similarity search (6 tools)
- SpatiaLite-Geospatial - Enterprise GIS with spatial analytics (7 tools)
- Vector-Index-Optimization - ANN search with k-means clustering (4 tools)
- Tool-Filtering - 🆕 v2.6.4: Selectively enable/disable tools for client compatibility
- MCP-Resources-and-Prompts - Intelligent workflow automation (7 resources + 7 prompts)
- Best-Practices - Usage patterns and recommendations
- Troubleshooting - Common issues and solutions
- Changelog - Version history and release notes
73 Total Tools Across 14 Categories (all tested and verified ✅):
- 🎯 JSON Helper Tools (6) - v2.6.0+: Parameter serialization, auto-normalization, JSONB binary storage
- 📊 Core Database (15) - Essential CRUD operations, schema management, and database administration
- 📝 Advanced Text Processing (8) - Regex, fuzzy matching, phonetic, similarity, normalization, validation
- 📈 Statistical Analysis (8) - Descriptive stats, percentiles, time series analysis
- 🔍 Full-Text Search (3) - FTS5 search, indexing, and content management
- 💾 Backup & Restore (5) - Database backup, restore, and migration tools
- ⚙️ PRAGMA Operations (5) - Database configuration and optimization
- 📋 Virtual Tables (6) - CSV import/export and virtual table management
- 📊 Enhanced Virtual Tables (4) - Smart CSV/JSON import with automatic schema inference
- 🧠 Semantic/Vector Search (6) - Embeddings and similarity search
- 🌍 Geospatial (SpatiaLite) (7) - Spatial data analysis and operations
- ⚡ Performance Optimization (4) - Vector indexing and ANN search optimization
- 🤖 MCP Resources (7) - Dynamic database meta-awareness resources
- 🔒 Security & Validation (5) - SQL injection protection and JSON validation
Reduce tool count for MCP clients with limits:
# Reduce to ~40 tools (Windsurf-compatible)
SQLITE_MCP_TOOL_FILTER="-vector,-stats,-spatial,-text"
# Read-only mode
SQLITE_MCP_TOOL_FILTER="-write_query,-create_table"
# Core + JSON only (~14 tools)
SQLITE_MCP_TOOL_FILTER="-fts,-vector,-virtual,-spatial,-text,-stats,-admin,-misc"See Tool Filtering for complete documentation.
Run instantly with Docker (creates a project-local database):
docker run -i --rm \
-v $(pwd):/workspace \
writenotenow/sqlite-mcp-server:latest \
--db-path /workspace/sqlite_mcp.dbWith tool filtering for Windsurf:
docker run -i --rm \
-e SQLITE_MCP_TOOL_FILTER="-vector,-stats,-spatial,-text" \
-v $(pwd):/workspace \
writenotenow/sqlite-mcp-server:latest \
--db-path /workspace/sqlite_mcp.db- 📋 Changelog - Version history and release notes
- 📰 v2.6.0 Release Article - JSON operations, auto-normalization, and enhanced security
- 🤖 AI-Powered Wiki Search - Search this documentation with natural language queries
- GitHub Repository - Source code and issues
- Docker Hub - Container images
- PyPI Package - Python package
- GitHub Gists - 9 practical examples and use cases
- Testing Guide - Comprehensive testing documentation
- Security Policy - Security guidelines and reporting
- Contributing - How to contribute to the project
- GitHub Releases - Version history
📝 Note: MCP Resources (7) and Prompts (7) are separate from the tool count - they provide metadata and workflow guidance but don't count toward the 73-tool limit.
🎛️ Tool Filtering: The server exposes 73 tools by default. Use SQLITE_MCP_TOOL_FILTER to reduce tool count for clients with limits. See Tool Filtering for details.