Find the right MCP server without the noise.
Browse practical, open-source Model Context Protocol servers by category, setup difficulty, publisher, and use case.
502 results
Showing 1-24 of 502
AWS MCP
awslabs
Manage AWS infrastructure — EC2 instances, S3 buckets, Lambda functions, CloudWatch logs, and IAM roles. Official Amazon build that gives AI assistants controlled access to your AWS environment for cloud resource management.
HexStrike AI MCP
0x4m4
Autonomous cybersecurity tool execution — runs 150-plus security tools for penetration testing, vulnerability scanning, and bug bounty automation. Designed for security researchers who want AI-assisted reconnaissance and automated security assessment workflows.
CodeGraphContext MCP
CodeGraphContext
Indexes your local codebase into a graph database for AI-powered code analysis. Maps relationships between files, functions, classes, and imports — then visualizes the graph so you understand complex codebases at a glance.
Appwrite MCP
appwrite
Manage Appwrite projects — databases, storage, users, functions, and authentication. Full backend management through AI for the open-source BaaS platform. Great for rapid prototyping and backend-as-a-service workflows.
ArangoDB MCP
arangodb
Multi-model database interactions for graph, document, and key-value queries. Query AQL, manage collections, and inspect graph relationships through AI-assisted database exploration across three data models.
BigQuery MCP
GoogleCloudPlatform
Run SQL queries on Google BigQuery, inspect datasets and tables, and analyze large-scale data warehouses through natural language. Built for data analysts and engineers working with petabyte-scale data.
Cassandra MCP
datastax
Query Apache Cassandra and ScyllaDB clusters — run CQL queries, inspect keyspaces and tables, manage indexes, and monitor cluster health. Designed for high-scale, distributed NoSQL workloads with tunable consistency.
Chroma MCP
chroma-core
Open-source embedding database for AI applications. Create collections, add documents with embeddings, and run semantic similarity searches. Purpose-built for RAG pipelines and LLM memory with automatic embedding management.
ClickHouse MCP
ClickHouse
Query ClickHouse databases for real-time analytics on large datasets. Designed for teams running OLAP workloads, event analytics, and time-series data at scale with columnar storage and vectorized execution.
CockroachDB MCP
cockroachdb
Manage CockroachDB clusters — run SQL queries, inspect database schemas, manage user permissions, and monitor distributed database health. Built for resilient, multi-region PostgreSQL-compatible deployments with automatic survivability.
Couchbase MCP
couchbase
Interact with Couchbase NoSQL databases — query documents, manage buckets, create indexes, and monitor cluster health. Supports N1QL queries through AI-assistant interaction for modern application backends.
Databricks MCP
databricks
Query Databricks SQL warehouses, explore Unity Catalog metadata, and run Spark SQL queries. Bridges the data lakehouse with AI-assisted analytics for data engineers and scientists working with big data infrastructure.
Cloudflare D1 MCP
cloudflare
Manage Cloudflare D1 serverless SQL databases — create databases, execute queries, run migrations, and inspect tables. Combines edge computing with SQLite-based persistence for globally distributed data access.
Drizzle ORM MCP
drizzle-team
Generate and run Drizzle ORM schema migrations, manage database relationships, and introspect existing databases. Streamlines TypeScript database workflows with AI-guided schema design for modern web applications.
DuckDB MCP
duckdb
In-process analytical SQL database for data analysis. Query CSV, Parquet, and JSON files directly using SQL without loading them into a separate database. Perfect for data scientists and analysts doing local data exploration.
DynamoDB MCP
awslabs
Query and manage Amazon DynamoDB tables — run Scan and Query operations, manage indexes, configure auto-scaling, and monitor read/write capacity. Built for serverless application teams operating at any scale.
FaunaDB MCP
fauna
Query Fauna document-relational databases using FQL. Manage collections, indexes, and documents with globally distributed, serverless database access through natural language. Built for edge and serverless architectures.
Firebase MCP
firebase
Manage Firebase services — Firestore, Realtime Database, Authentication, and Storage. Query documents, manage user authentication, and configure security rules through AI. Comprehensive Google Firebase platform management.
Firestore MCP
googleapis
Google Firestore NoSQL document database operations — create and query collections, manage documents, set up real-time listeners, and configure composite indexes for scalable applications with offline support.
InfluxDB MCP
influxdata
Query InfluxDB time-series databases — write and read measurements, manage buckets and retention policies, and run Flux queries. Essential for IoT monitoring, metrics collection, and observability pipelines handling high-cardinality data.
MariaDB MCP
mariadb
Full MariaDB database management — query tables, manage users, configure replication, and monitor performance. Enterprise-grade database administration through AI with MySQL-compatible syntax for production environments.
Memgraph MCP
memgraph
Real-time graph database querying using Cypher. Analyze connected data, detect patterns, and traverse graph relationships with millisecond latency for streaming and static datasets in fraud detection and recommendation engines.
Milvus MCP
milvus-io
Query and manage Milvus vector databases for AI applications. Create collections, insert vectors, run similarity searches, and manage indexes — essential for RAG and semantic search pipelines at production scale.
MongoDB MCP
kiliczsh
Interact with MongoDB databases through AI — run aggregations, inspect collections, manage indexes, and query documents using natural language instead of writing complex MongoDB queries by hand.
What MCP solves
MCP gives AI assistants one shared way to connect with databases, repositories, APIs, files, and internal tools without rebuilding every integration from scratch.
Local-first by default
Most servers run as local processes through stdio, so credentials and private data can stay on the user's machine unless a remote service is explicitly involved.
Built for workflow depth
The useful servers expose clear tools, resources, and prompts so the assistant can inspect context, take action, and return structured results.
Pick the connector before the workflow breaks
MCP is useful when the server does one job well.
The best MCP servers are boring in the right way: clear setup, narrow permissions, predictable tool names, and docs that tell you what happens to your data. That is what this page is meant to help you spot before you wire a server into Claude, Cursor, VS Code, or a local agent stack.
Search by the thing you need the assistant to touch first. Repository? Start with version control. Customer data? Look at database servers and read the auth notes twice. Browser task? Prefer automation servers that expose screenshots, navigation, and extraction as separate actions.
A quick pre-install check
Confirm the transport
Most desktop workflows use stdio. Remote HTTP servers are better for shared team infrastructure, but they need tighter auth and logging.
Read the tool list
Good servers expose small, named actions. Vague tools like “run anything” are powerful, but risky.
Check credential handling
Look for environment variable examples, scoped API keys, and a clear story for secrets.
Test with a throwaway task
Before connecting production data, ask the assistant to perform a harmless read-only action and inspect the result.
Common MCP paths people search for
Where MCP shines
Use MCP when the assistant needs live context: read a repo, query a database, inspect a ticket, fetch docs, or drive a browser. For one-off text generation, you probably do not need a server at all.