MCP Servers for Databases, Warehouses and Vector Stores
Database servers are the highest-leverage and highest-risk group on the site. An assistant that can read your schema stops guessing at column names, but the same connection can run a statement you did not intend, so most of these guides spend real time on read-only credentials.
Two costs are easy to miss. Warehouse servers bill per query scanned, which the BigQuery and Snowflake guides both address, and vector servers may download an embedding model on first run before they answer anything.
19 guides in this collection.
Relational and serverless Postgres
PostgreSQL connection setup
Set up the PostgreSQL MCP server in Cursor or Windsurf and let your AI query tables, write SQL, explain schemas, and generate migrations — without copy-pasting data. Full setup guide for 2026.
Supabase for Claude and Cursor
How to connect your Supabase database to Claude Desktop, Cursor, and Windsurf using the official Supabase MCP server. Query tables, run migrations, and manage your project with AI.
Neon branching and migrations
Neon MCP server setup for Cursor IDE: generate an API key, configure @neondatabase/mcp-server-neon, and let your AI create database branches, run SQL, and manage projects without touching production.
Prisma Postgres provisioning
Prisma MCP server Cursor IDE setup 2026: connect to mcp.prisma.io, authorize against Prisma Console, then provision databases, run SQL, and manage backups without leaving Cursor.
MySQL queries from chat
Connect MySQL to Cursor IDE with an MCP server so your AI can read schemas, write SQL, and debug slow queries without copy-pasting table structures. Includes a read-only user setup and mcp.json config.
PlanetScale branches and query safeguards
PlanetScale MCP server Cursor IDE setup 2026: one url entry, OAuth in the browser, then inspect branches, run queries, and read Insights from chat with built-in destructive-query blocks.
Turso read-only SQLite and libSQL
Connect a Turso database to Cursor IDE via the community mcp-turso server: get your database URL and auth token from the Turso CLI, configure mcp.json, and run read-only SELECT queries from chat.
Warehouses and analytics
BigQuery datasets and SQL
Connect Google BigQuery to Cursor IDE using MCP. Run SQL, explore datasets, and analyze billions of rows from Composer — full setup guide with IAM config.
Snowflake with warehouse cost guardrails
Connect Snowflake's managed MCP server to Cursor IDE: generate a Programmatic Access Token, point mcp.json at your account's MCP endpoint, and query warehouses safely — including why the old community server is being retired and how to stop an AI from running up compute cost.
ClickHouse cloud and self-hosted
Connect ClickHouse to Cursor IDE two ways: ClickHouse Cloud's OAuth-based remote MCP server, or the open-source mcp-clickhouse server via uv for self-hosted clusters. Runs read-only by default — writes need an explicit opt-in flag.
Google Cloud MCP Toolbox for Databases
Google Cloud MCP server setup for Cursor IDE, 2026: install MCP Toolbox for Databases, define sources and tools in tools.yaml, and query Cloud SQL, AlloyDB, Spanner, or BigQuery from chat.
Document, key-value and vector stores
MongoDB collections and schemas
Connect MongoDB to Cursor IDE via MCP and query your collections, inspect schemas, and debug data issues directly while writing code. Works with Atlas and self-hosted MongoDB.
DynamoDB tables
DynamoDB MCP server Cursor IDE setup 2026: install awslabs.dynamodb-mcp-server via uvx, configure AWS_PROFILE and AWS_REGION, and get real data-modeling and cost-estimation tools in chat.
Convex deployments
Set up the Convex MCP server in Cursor IDE — it ships inside the Convex CLI, so there's no separate package. Inspect tables, run functions, and read logs without opening the dashboard.
Redis keys and data structures
Connect Redis to Claude, Cursor, and VS Code with MCP. Query keys, inspect data structures, run commands, and search vector indexes from your AI assistant.
Upstash Redis, QStash and Workflow
Upstash MCP server Cursor IDE setup 2026: one npx command with your email and API key, then manage Redis databases, QStash, and Workflow runs directly from Cursor chat.
Elasticsearch indices and slow searches
Set up an Elasticsearch MCP server in Cursor IDE with a scoped API key, mcp.json config, and real prompts for inspecting mappings and debugging slow queries — plus why you should never point it at production with an admin key.
Qdrant memory store via uvx
Connect Qdrant's official MCP server to Cursor IDE: install with uvx (not npx), set QDRANT_URL or QDRANT_LOCAL_PATH plus COLLECTION_NAME, and store or retrieve memories with the qdrant-store and qdrant-find tools — including why this one needs Python's uv, not Node.
Pinecone vector index search
Connect Pinecone's official developer MCP server to Cursor IDE: add a .cursor/mcp.json entry with your PINECONE_API_KEY, then search docs, list indexes, and upsert records from chat — plus why it only works with integrated-inference indexes, not raw vector search.
Other kinds of MCP server
Prefer to scan the whole library at once? The complete article index lists every guide by publication date, and both hub views sit side by side on the guides page.